Articles

Cetuximab, docetaxel, and cisplatin versus platinum,

fluorouracil, and cetuximab as first-line treatment in
patients with recurrent or metastatic head and neck
squamous-cell carcinoma (GORTEC 2014-01 TPExtreme):
a multicentre, open-label, randomised, phase 2 trial
Joël Guigay*, Anne Aupérin*, Jérôme Fayette, Esma Saada-Bouzid, Cédrik Lafond, Miren Taberna, Lionnel Geoffrois, Laurent Martin,
Olivier Capitain, Didier Cupissol, Hélène Castanie, Damien Vansteene, Philippe Schafhausen, Alison Johnson, Caroline Even, Christian Sire,
Sophie Duplomb, Camille Evrard, Jean-Pierre Delord, Brigitte Laguerre, Sylvie Zanetta, Cécile Chevassus-Clément, Aldéric Fraslin, Fanny Louat,
Laura Sinigaglia, Ulrich Keilholz†, Jean Bourhis†, Ricard Mesia†, on behalf of the GORTEC, AIO, TTCC, and UniCancer Head and Neck groups‡

Summary
Background Results from a phase 2 trial of the TPEx chemotherapy regimen (docetaxel–platinum–cetuximab) showed Lancet Oncol 2021
promising results, with a median overall survival of 14·0 months in first-line recurrent or metastatic head and neck Published Online
squamous-cell carcinoma (HNSCC). We therefore aimed to compare the efficacy and safety of the TPEx regimen with March 5, 2021
https://doi.org/10.1016/
the standard of care EXTREME regimen (platinum–fluorouracil–cetuximab) in this setting.
S1470-2045(20)30755-5
See Online/Comment
Methods This was a multicentre, open-label, randomised, phase 2 trial, done in 68 centres (cancer centres, university https://doi.org/10.1016/
and general hospitals, and private clinics) in France, Spain, and Germany. Eligible patients were aged 18–70 years S1470-2045(21)00121-2
with histologically confirmed recurrent or metastatic HNSCC unsuitable for curative treatment; had at least one *Joint first authors
measurable lesion according to Response Evaluation Criteria in Solid Tumors version 1.1; and had an Eastern †Joint last authors
Cooperative Oncology Group (ECOG) performance status of 1 or less. Participants were randomly assigned (1:1) using ‡Members are listed in the
the TenAlea website by investigators or delegated clinical research associates to the TPEx regimen or the EXTREME appendix pp 2–3
regimen, with minimisation by ECOG performance status, type of disease evolution, previous cetuximab treatment, Department of Medical
and country. The TPEx regimen consisted of docetaxel 75 mg/m² and cisplatin 75 mg/m², both intravenously on Oncology and Research
(Prof J Guigay MD) and
day 1, and cetuximab on days 1, 8, and 15 (intravenously 400 mg/m² on day 1 of cycle 1 and 250 mg/m² weekly
Department of Medical
subsequently). Four cycles were repeated every 21 days with systematic granulocyte colony-stimulating factor (G-CSF) Oncology (E Saada-Bouzid MD),
support at each cycle. In case of disease control after four cycles, intravenous cetuximab 500 mg/m² was continued Centre Antoine Lacassagne,
every 2 weeks as maintenance therapy until progression or unacceptable toxicity. The EXTREME regimen consisted FHU Oncoage, University
Côte d’Azur, Nice, France;
of fluorouracil 4000 mg/m² on day 1–4, cisplatin 100 mg/m² on day 1, and cetuximab on days 1, 8, and 15 (400 mg/m²
Biostatistic and Epidemiology
on day 1 of cycle 1 and 250 mg/m² weekly subsequently) all delivered intravenously. Six cycles were delivered every Unit, Gustave Roussy, Oncostat
21 days followed by weekly 250 mg/m² cetuximab as maintenance therapy in case of disease control. G-CSF support 1018 INSERM, labeled Ligue
was not mandatory per the protocol in the EXTREME regimen. The primary endpoint was overall survival in the Contre le Cancer, University
Paris-Saclay, Villejuif, France
intention-to-treat population; safety was analysed in all patients who received at least one dose of chemotherapy or
(A Aupérin MD,
cetuximab. Enrolment is closed and this is the final analysis. This study is registered at ClinicalTrials.gov, C Chevassus-Clement MSc,
NCT02268695. A Fraslin MSc); Department of
Medical Oncology, University
of Lyon, Centre Léon Bérard,
Findings Between Oct 10, 2014, and Nov 29, 2017, 541 patients were enrolled and randomly assigned to the Lyon, France (J Fayette MD);
two treatment regimens (271 to TPEx, 270 to EXTREME). Two patients in the TPEx group had major deviations in Department of Medical
consent forms and were not included in the final analysis. Median follow-up was 34·4 months (IQR 26·6–44·8) in Oncology, ILC Centre Jean
the TPEx group and 30·2 months (25·5–45·3) in the EXTREME group. At data cutoff, 209 patients had died in the Bernard/Clinique Victor Hugo,
Le Mans, France (C Lafond MD);
TPEx group and 218 had died in the EXTREME group. Overall survival did not differ significantly between the groups Department of Medical
(median 14·5 months [95% CI 12·5–15·7] in the TPEx group and 13·4 months [12·2–15·4] in the EXTREME group; Oncology, Institut Català de
hazard ratio 0·89 [95% CI 0·74–1·08]; p=0·23). 214 (81%) of 263 patients in the TPEx group versus 246 (93%) of 265 Oncologia, L’Hospitalet de
patients in the EXTREME group had grade 3 or worse adverse events during chemotherapy (p<0·0001). In the TPEx Llobregat, Barcelona, Spain
(M Taberna MD, R Mesia MD);
group, 118 (45%) of 263 patients had at least one serious adverse event versus 143 (54%) of 265 patients in the Department of Medical
EXTREME group. 16 patients in the TPEx group and 21 in the EXTREME group died in association with adverse Oncology, Institut de
events, including seven patients in each group who had fatal infections (including febrile neutropenia). Eight deaths Cancérologie de Lorraine,
in the TPEx group and 11 deaths in the EXTREME group were assessed as treatment related, most frequently sepsis Nancy, France (L Geoffrois MD);
Department of Medical
or septic shock (four in each treatment group). Oncology, Clinique des
Ormeaux, Le Havre, France
Interpretation Although the trial did not meet its primary endpoint, with no significant improvement in overall (L Martin MD); Department of
survival with TPEx versus EXTREME, the TPEx regimen had a favourable safety profile. The TPEx regimen could Medical Oncology, Institut de

www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5 1

 Articles

Cancérologie de l’Ouest provide an alternative to standard of care with the EXTREME regimen in the first-line treatment of patients with
Paul Papin, Angers, France recurrent or metastatic HNSCC, especially for those who might not be good candidates for up-front pembrolizumab
(O Capitain MD); Department of
Medical Oncology, Institut du
treatment.
Cancer de Montpellier,
Montpellier, France Funding Merck Santé and Chugai Pharma.
(D Cupissol MD); Department of
Medical Oncology, Hôpital
Privé le Confluent, Sainte
Copyright © 2021 Elsevier Ltd. All rights reserved.
Catherine de Sienne, Nantes,
France (H Castanie MD); Introduction do so.4–7 With the advent of immunotherapy, the standard
Department of Medical When this trial was initiated in 2014, the EXTREME of care EXTREME regimen was replaced in 2020 by
Oncology, Institut de
Cancérologie de l’Ouest René
regimen (platinum, fluorouracil, and cetuximab followed pembrolizumab, which improves overall survival but not
Gauducheau, Nantes, France by weekly 250 mg/m² cetuximab maintenance) was progression-free survival, alone or in combination with
(D Vansteene MD); Department considered the first-line standard treatment option in platinum and fluorouracil for patients with a PD-L1
of Medical Oncology, patients with recurrent or metastatic head and neck combined positive score of 1 or more.8,9 However, in
Universitätsklinikum Hamburg
Eppendorf Hubertus Wald
squamous-cell carcinoma (HNSCC) not suitable for some countries, such as the USA, pembrolizumab has
Tumorzentrum, Hamburg, locoregional curative treatment.1,2 Treatment goals in this also been approved for patients with a combined positive
Germany (P Schafhausen MD); setting are to prolong survival and delay progression, score of less than 1, even though subgroup analyses do
Department of Medical
while maintaining quality of life.3 However, survival not support benefit in these patients.8–10
Oncology, Centre François
Baclesse, Caen, France results with the EXTREME regimen were far from On the basis of preclinical data suggesting a synergistic
(A Johnson MD); Department of satisfactory, with median overall survival of 10·1 months effect of taxanes and cetuximab,11 combinations of taxanes
Head and Neck Oncology, and median progression-free survival of 5·6 months.1 and cetuximab with or without platinum have been
Gustave Roussy, Villejuif,
In this context, the development of new therapies is studied with promising antitumour efficacy.12–21 The
France (C Even MD);
Department of Medical important to improve these outcomes. Several targeted GORTEC 2008-03 TPEx phase 2 study evaluating the
Oncology, Centre Hospitalier therapies, tested alone or in combination, have failed to TPEx regimen (four cycles of docetaxel in combination
de Bretagne Sud, Lorient,
France (C Sire MD); Department
of Medical Oncology, HCL,
Research in context
Hôpital de la Croix Rousse,
Lyon, France (S Duplomb MD); Evidence before this study a taxane offered several advantages: shorter treatment
Department of Medical
Oncology, CHU, Université de
We searched prospective clinical trial publications, published in duration, easier delivery in daily practice, and fewer
Poitiers, Poitiers, France English and indexed in PubMed from Sept 1, 2008, to contraindications than the standard fluorouracil–cisplatin
(C Evrard MD); Department of Sept 1, 2020, for the title or abstract terms “head and neck” and combination used in the EXTREME regimen.
Medical Oncology, Institut “carcinoma”, or “cancer” and “first-line” and “recurrent”, or
Claudius Regaud, Toulouse, Added value of this study
France (Prof J-P Delord MD);
“metastatic” and “randomised”. The search returned
This randomised trial assessed the efficacy and safety of the
Department of Medical 47 publications, most of which used chemotherapy–cetuximab-
TPEx regimen in first-line treatment of patients with recurrent
Oncology, Centre Eugène based combinations. When our trial was initiated in 2014,
Marquis, Rennes, France or metastatic HNSCC compared with the EXTREME regimen.
the EXTREME regimen (platinum, fluorouracil, and cetuximab
(B Laguerre MD); Department of We showed good survival results of the TPEx regimen, similar to
Medical Oncology, Centre
followed by weekly cetuximab maintenance) was considered
those observed in the initial phase 2 trial, but did not show a
Georges-François Leclerc, the first-line standard option in patients with recurrent or
significant improvement in overall survival compared with the
Dijon, France (S Zanetta MD); metastatic head and neck squamous-cell carcinoma (HNSCC)
GORTEC, Tours, France EXTREME regimen. Compared with the EXTREME regimen, the
not suitable for locoregional curative treatment. However,
(F Louat PhD, L Sinigaglia MSc, TPEx regimen included a shorter course of chemotherapy (four
Prof J Bourhis MD); Charité survival results were far from satisfactory, and new therapies
cycles instead of six cycles), a less frequent cetuximab
Comprehensive Cancer Center, are needed to improve outcomes. With the advent of
maintenance treatment schedule (every 2 weeks instead of
Charité, Berlin, Germany immunotherapy, the first-line standard of care EXTREME
(Prof U Keilholz MD) weekly doses), and was better tolerated and provided better
regimen was replaced by pembrolizumab, which improves
Correspondence to: quality of life.
overall survival but not progression-free survival, alone or in
Prof Joël Guigay, Department of
Medical Oncology, Centre
combination with platinum and fluorouracil for patients with a Implications of all the available evidence
Antoine Lacassagne, PD-L1 combined positive score of 1 or more. Based on Although our trial did not meet its primary endpoint, the results
06189 Nice Cedex 2, France preclinical data suggesting a synergistic effect of taxanes and are informative and could potentially change practice, because
joel.guigay@nice.unicancer.fr
cetuximab, combinations of taxanes with or without platinum the TPEx regimen could be an alternative to the EXTREME
See Online for appendix and cetuximab have been studied with promising antitumour regimen in first-line treatment of patients with recurrent or
efficacy. The GORTEC 2008-03 TPEx phase 2 study evaluating metastatic HNSCC, especially for those with a negative PD-L1
four cycles of docetaxel in combination with cisplatin and combined positive score, those who might not be good
cetuximab followed by cetuximab maintenance every 2 weeks candidates for up-front pembrolizumab because of
(the TPEx regimen) showed promising results with a median immunologically relevant comorbidities, patients with high
overall survival of 14·0 months (which compared favourably tumour burden or symptoms that mean a rapid response is a key
with the EXTREME regimen). The substitution of fluorouracil by treatment goal, or patients with contraindication to fluorouracil.

2 www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5

 Articles

with cisplatin and cetuximab followed by 500 mg/m² The study was done in accordance with Good Clinical
cetuximab maintenance every 2 weeks) showed promising Practice Guidelines and the Declaration of Helsinki and
results, with a median overall survival of 14·0 months19 was approved by competent authorities and ethics
(compared with the EXTREME regimen’s median overall committees in July, 2014 (France), October, 2014 (Spain),
survival of about 10 months1,8). The substitution of and May, 2015 (Germany). An international, independent
fluorouracil by a taxane offered several advantages: data and safety monitoring committee, which included
shorter treatment duration, easier delivery in daily two oncologists and one statistician, monitored progress
practice, and fewer contra­indi­cations than fluorouracil and interim analysis reports.
(which is contraindicated in patients with conditions
such as dihydropyrimidine dehydrogenase deficiency and Randomisation and masking
ischaemic cardiac disease). Patients were randomly assigned (1:1) using the TenAlea For the TenAlea system see
Based on this rationale, our aim was to assess the website by investigators or delegated clinical research https://prod.tenalea.net/igr/dm/

efficacy and safety of the TPEx regimen in the first-line associates to receive the TPEx regimen or the EXTREME
treatment of patients with recurrent or metastatic regimen with minimisation by ECOG performance
HNSCC compared with the standard of care EXTREME status (0 vs 1), type of disease evolution (locoregional
regimen. relapse alone vs metastatic disease), previous cetuximab
treatment (no vs yes), and country (France vs Germany vs
Methods Spain). To avoid deterministic minimisation and assure
Study design and participants allocation conceal­ment, the treatment that minimised
This was a multicentre, open-label, randomised, phase 2 the imbalance was assigned with a probability of 0·80.
trial, done in 68 centres (cancer centres, university and Minimisation parameters were defined by the Gustave
general hospitals, and private clinics) in France, Spain, Roussy Biostatistics Unit (Villejuif, France) in the
and Germany (appendix pp 2–3). Eligible patients were TenAlea system. Physicians and patients were not
aged 18–70 years; had histologically confirmed masked to treatment group.
squamous-cell carcinoma of the oral cavity, oropharynx,
hypopharynx, or larynx, with metastases or recurrence Procedures
not suitable for locoregional curative treatment; had at The TPEx regimen (appendix p 4) consisted of docetaxel
least one measurable lesion according to Response 75 mg/m² as a 1 h intravenous infusion on day 1, cisplatin
Evaluation Criteria in Solid Tumors version 1.1 75 mg/m² as a 1 h intravenous infusion on day 1, and
(RECIST 1.1); had an Eastern Cooperative Oncology cetuximab on days 1, 8, and 15 (400 mg/m² at 5 mg/min
Group (ECOG) perfor­mance status of 1 or less; were maximum speed intravenous infusion on day 1 of cycle 1
eligible to receive cisplatin; had clearance of creatinine and 250 mg/m² at 10 mg/min maximum speed intra­
more than 60 mL/min per 1·73 m² (by the Modification venous infusion weekly on subsequent admin­istrations).
of Diet in Renal Disease method); absolute neutrophil Cetuximab infusion ended at least 1 h before the start of
count of more than 1·5 × 10⁹ cells per L; platelet count of cisplatin followed by docetaxel infusion. Four cycles were
more than 100 × 10⁹ per L; haemoglobin concentration repeated every 21 days with systematic granulocyte colony-
of at least 9·5 g/dL; bilirubin concentration at or below stimulating factor (G-CSF; lenograstim was recom­
the upper limit of normal (ULN); aspartate amino­ mended, but filgrastim was also authorised according to
transferase and alanine amino­ transferase concen­ local guidelines in investigational sites) support at each
trations of less than 1·5 times the ULN; and alkaline cycle. In case of disease control after four cycles,
phosphatase concentration of less than 2·5 times the intravenous cetuximab 500 mg/m² was continued every
ULN. Exclusion criteria included: previous systemic 2 weeks as maintenance therapy until progression or
chemotherapy for HNSCC (except if admin­istered as unacceptable toxicity.
part of a multimodal treatment for locally advanced The EXTREME regimen (appendix p 4) consisted of
disease more than 6 months before study entry); surgery fluorouracil 4000 mg/m² as a 96 h continuous intra­
or radiotherapy within the previous 6 weeks; previous venous infusion on days 1–4, cisplatin 100 mg/m² as a 1 h
dose of cisplatin more than 300 mg/m²; treatment with intravenous infusion on day 1, and cetuximab on days 1,
EGFR-targeting therapy within the previous 12 months; 8, and 15 (400 mg/m² at 5 mg/min maximum speed
clinically significant cardiovascular disease; other intravenous infusion on day 1 of cycle 1 and 250 mg/m²
malignancies within 5 years before randomisation, with at 10 mg/min maximum speed intravenous infusion
the exception of adequately treated basal skin cancer weekly on subsequent administrations). Six cycles were
and carcinoma in situ of the cervix; active infection delivered every 21 days followed by weekly 250 mg/m²
requiring intravenous antibiotic drugs; tuberculosis cetuximab as maintenance therapy in case of disease
infection; and HIV infection (complete list of inclusion control. According to the summary of product For the protocol see https://
www.gortec.net/protocoles/
and exclusion criteria are available in the protocol). All characteristics of cetuximab and standard recommen­
TPExtreme_Protocole_
patients gave written informed consent before any study dations, G-CSF support was not mandatory per protocol V5.0_22_12_2016_version_
procedure. in the EXTREME group. finale.pdf

www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5 3

 Articles

In both groups, in case of toxicity prohibiting chemo­ Outcomes

therapy continuation, maintenance with cetuximab could The primary endpoint was overall survival, defined as the
be started after two cycles of chemotherapy if patients interval between randomisation and death from any cause.
had stable disease or an objective response. Secondary endpoints were progression-free survival,
Tumour response was assessed every 8 weeks after the defined as the interval between randomisation and first
start of treatment by CT scan for the neck, chest, and disease progression (investigator assessed according to
abdomen or MRI for the neck until disease progression. RECIST 1.1) or death, whichever occurred first; time to
Determination of human papillomavirus (HPV) status progression, defined as the minimum time from
for patients with oropharyngeal primary tumour was randomisation to progression, with censoring in case of
analysed centrally by chromogenic in-situ hybridisation death from a cause other than cancer without previous
(CISH) for DNA of HPV 16, HPV 18, and HPV 33. HPV progression (investigator assessed); objective response rate
DNA CISH has higher specificity than p16 expression (complete or partial response according to RECIST 1.1) at
by immunohistochemistry, with the trade-off of lower week 12 evaluated by independent central review;
sensitivity in oropharyngeal primaries. Therefore, HPV investigator-assessed best overall response according to
DNA CISH was favoured for the study. RECIST 1.1 during treatment; QOL as assessed by the
Health-related quality of life (QOL) was assessed using QLQ-C30 questionnaire; compliance with chemotherapy
the European Organisation for Research and Treatment and cetuximab; and safety. A medico-economic study is
of Cancer (EORTC) Quality of Life Questionnaire core ongoing, the results of which will be reported elsewhere.
module QLQ-C30 at baseline, and at weeks 12, 18, and 26.
Adverse event monitoring was done weekly during the Statistical analysis
chemotherapy phase and every 2 weeks during Sample size was calculated for detecting a hazard ratio
maintenance therapy using physical examination, check (HR) between the TPEx regimen and EXTREME regimen
of vital signs, and blood sampling for haematological and groups of 0·72 for death, which corresponds to an
biochemistry assessments. Adverse events were collected increase of median overall survival time from 10·1 months
and graded according to the National Cancer Institute’s to 14·0 months. Under the assumption that deaths follow
Common Terminology Criteria for Adverse Events an exponential distribution, and at a two-sided level of
(NCI-CTCAE) version 4.03. statistical significance of 0·05, 374 deaths would provide
Occurrence of a severe infusion-related reaction 88% power, which would be expected to occur out of a
required immediate and permanent discontinuation of total of 540 patients. Initially, the trial was designed with
cetuximab therapy. A grade 1, mild, transient reaction 80% power (that required 295 deaths and 416 patients),
(transient flushing or rash or drug fever <38°C) required which was increased to 88% in December, 2016, because
a decrease of the cetuximab infusion rate by 50% and close the rate of enrolment was high so a power increase was
monitoring for any worsening; the infusion rate was authorised without extending the accrual period. One
further decreased if the reaction persisted. Modifications interim futility analysis of overall survival was planned
to the doses of chemotherapy were based on toxicity when around 50% of total expected deaths had occurred.
observed during the previous cycle. In case of neutro­ The z-score futility boundary was constructed using the
penia, thrombocytopenia grade 3–4, or both, without spending function of Lan-DeMets and was non-binding.
fever, docetaxel was stopped until normalisation. Chemo­ Time-to-event endpoints (overall survival, progression-
therapy was stopped for 7 days in case of fever higher free survival, and time to progression) were estimated
than 38·5°C or a grade 3–4 adverse event; in these cases, using the Kaplan-Meier method. The 95% CIs of the
further doses of docetaxel were reduced by 20%. Only point estimates at 12, 24, and 36 months of overall
two dose modifications were permitted per patient. Study survival and progression-free survival were calculated
treatment was definitively stopped in case of absence of with the Rothman method. HRs were estimated using
normalisation at day 36 despite a previous decrease of the Cox model. Crude HRs and HRs from the Cox model
docetaxel doses. Non-haematological chemotherapy- stratified for country and adjusted for the other
related toxicities had to be resolved to grade 0 (excluding minimisation factors (ECOG performance status, type of
skin reactions, paronychia, alopecia, fatigue, ototoxicity, disease evolution, and previous cetuximab treatment) are
or neurotoxicity, which must have resolved to grade ≤2). reported. The proportional hazards assumption for the
In case of cisplatin-related nephrotoxicity of grade 1 or Cox models was assessed by plotting the log-negative-log
worse, ototoxicity of grade 3 or worse, or neurotoxicity of of the Kaplan-Meier estimates of the survival function
grade 3 or worse, cisplatin was replaced by intravenous versus the log of time and there was no indication that
carboplatin with an area under the curve of 5 in the this assumption had been violated for overall survival or
following cycles, with a maximum dose of 750 mg per progression-free survival. The rate of objective response
cycle. Details on permitted dose modifications and at week 12 by independent central review was estimated
adverse event monitoring are provided in the protocol. as the number of patients with complete or partial
The only criterion for patient removal from the study was response among all patients with imaging provided by
consent withdrawal. centres, even those with imaging that was not evaluable.

4 www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5

 Articles

The rate of best overall response by investigator

assessment was estimated as the number of patients 541 patients enrolled and randomly assigned

with complete or partial response at any time during

treatment among all patients, even those who were not
evaluable or not evaluated. Comparisons of these rates 271 assigned to the TPEx regimen 270 assigned to the EXTREME regimen
between treatment groups were done using a χ² test.
The proportion of patients receiving the planned
number of chemotherapy cycles and the proportion of 2 major deviations in informed consent
patients receiving the planned number of cetuximab
administrations were compared between treatment 263 completed or discontinued treatment 265 completed or discontinued treatment
groups using a χ² test. (including 2 who received one (including 2 who received the TPEx
Efficacy was analysed in the intention-to-treat administration of cetuximab only) regimen and 1 who received one
6 did not receive any study treatment administration of cetuximab only)
population according to the randomly assigned treatment 1 had symptomatic deterioration 5 did not receive any study treatment
group. Safety was analysed in all patients who received at 1 died due to cancer 2 had symptomatic deterioration
1 had disease progression 1 died due to cancer
least one dose of chemotherapy or cetuximab. The 2 had adverse events (1 hepatic function 2 had adverse events (1 creatinine
protocol planned to use the χ² test to compare between disorder, 1 pneumonia) increase, 1 hearing impaired)
the two groups the distribution of patients according to 1 withdrew informed consent

their worst adverse events into three categories (no

adverse events, highest grade 1–2, highest grade ≥3). The 12 found to be ineligible after central review 10 found to be ineligible after central review
scores of the different scales of the EORTC QLQ-C30 3 previous dose of cisplatin >300 mg/m² 2 previous dose of cisplatin >300 mg/m²
questionnaire were compared between the two groups 2 liver enzyme increased higher than 1 hypoacousia after previous cisplatin
required 1 renal function worse than required
using a mixed model for repeated measures of QOL and 1 ECOG PS 2 2 <6 months between end of previous
taking into account the baseline value before treatment. 2 <6 months between end of previous systemic chemotherapy for head and
systemic chemotherapy for head and neck cancer and inclusion in trial
Exploratory prespecified subgroup analyses of efficacy neck cancer and inclusion in trial 2 no primary tumour
outcomes by baseline characteristics (sex; age; ECOG 2 no primary tumour 1 cerebral metastasis
performance status; type of disease evolution; tumour 1 cerebral metastasis 1 inclusion for initial treatment of
1 previous lung cancer within 5 years locoregional tumor without distant
location; and, in patients with oropharyngeal carcinoma, 2 lost to follow-up before 12 months metastasis
HPV DNA status) were done using the Cox model, 3 lost to follow-up before 12 months
including treatment group, baseline characteristic, and
term for interaction between treatment group and
269 included in intention-to-treat analysis 270 included in intention-to-treat analysis
baseline characteristic. Several post-hoc analyses were 263 included in safety analysis 265 included in safety analysis
done: a sensitivity analysis of the primary endpoint
excluding ineligible patients; a multivariable prognostic Figure 1: Trial profile
analysis of overall survival and progression-free survival; ECOG PS=Eastern Cooperative Oncology Group performance status. EXTREME=cisplatin, fluorouracil,
a comparison of overall survival between patients in and cetuximab. TPEx=docetaxel, cisplatin, and cetuximab.
second-line treatment by immunotherapy and by
chemotherapy; and, in the EXTREME regimen group, Results
comparison of the week 12 objective response rate and Between Oct 10, 2014, and Nov 29, 2017, 541 patients were
the best objective response rate, and an analysis of the enrolled and randomly assigned to the two treatment
association between G-CSF administration and overall groups (TPEx regimen n=271, EXTREME regimen n=270).
survival. The multivariable prognostic analysis studied Major deviations in informed consent forms were noted
initial patient and tumour characteristics (age, sex, for two patients in the TPEx group who were consequently
ECOG performance status, tobacco consumption, removed from the study after starting treatment and were
HPV DNA status, type of disease evolution, and previous not included in the intention-to-treat analysis. 539 patients
cetuximab use) using the Cox model stratified for were included in the primary and secondary analyses
country. (figure 1). Deviations from the eligibility criteria occurred
All p values were two-sided at a significance level of 0·05. in 22 patients; these patients were found to be ineligible
Only the comparison of overall survival between the on central review after randomisation but were still
groups is considered a confirmatory statistical test; all analysed as per the intention-to-treat principle (figure 1).
other statistical tests are supportive or of exploratory Baseline characteristics are shown in table 1.
nature. Analyses were done with SAS (version 9.4). This 525 (97%) of 539 patients received at least one course of
study is registered with ClinicalTrials.gov, NCT02268695. chemotherapy (figure 1). Two patients allocated to the
EXTREME group received the TPEx regimen by mistake.
Role of the funding source These patients were included in the EXTREME group in
The funders of the study had no role in study design, data all analyses.
collection, data analysis, data interpretation, or writing of An interim futility analysis was done on April 10, 2017,
the report. when 174 deaths had occurred (ie, 46·5% of the planned

www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5 5

 Articles

death was 0·91 (95% CI 0·67–1·22). The futility boundary

TPEx regimen EXTREME regimen
group (n=269) group (n=270) was not crossed and the independent data safety
monitoring committee recommended the trial to continue
Sex
after this analysis.
Female 29 (11%) 39 (14%)
At final analysis (data cutoff Nov 29, 2019), median
Male 240 (89%) 231 (86%)
follow-up was 34·4 months (IQR 26·6–44·8) in the TPEx
Age, years
group and 30·2 months (25·5–45·3) in the EXTREME
Median 60 (55–64) 60 (55–63)
group. At data cutoff, 427 patients had died: 209 in the
≥65 56 (21%) 47 (17%)
TPEx group and 218 in the EXTREME group (deaths
ECOG performance status
related to cancer n=161 in the TPEx group vs n=173 in the
0 86 (32%) 86 (32%)
EXTREME group; deaths from unknown causes n=17 vs
1 183 (68%) 184 (68%) n=19; deaths from other causes n=31 vs n=26). Overall
Regular alcohol consumption survival did not differ significantly between the groups
No 58 (22%) 77 (29%) (median overall survival 14·5 months [95% CI 12·5–15·7]
Yes (current or past) 206 (77%) 193 (71%) in the TPEx group vs 13·4 months [12·2–15·4] in the
Missing data 5 (2%) 0 EXTREME group; HR 0·89 [95% CI 0·74–1·08], p=0·23
Smoker [figure 2A]; adjusted HR 0·91 [0·75–1·10]).
No 14 (5%) 27 (10%) At data cutoff, there were 503 disease progression events
Yes (current or former) 255 (95%) 243 (90%) or deaths across both groups: 248 in the TPEx group
Primary tumour site (225 progression events and 23 deaths) and 255 in the
Oropharynx 120 (45%) 96 (36%) EXTREME group (231 progression events and 24 deaths).
Oral cavity 57 (21%) 52 (19%) Progression-free survival did not differ signifi­ cantly
Hypopharynx 54 (20%) 63 (23%) between the groups (figure 2B); adjusted HR 0·90 (95%
Larynx 34 (13%) 57 (21%) CI 0·75–1·07). Time to progression also did not differ
Multiple locations* 3 (1%) 0 significantly between the two groups (appendix p 7).
Other 0 1 (<1%) According to the independent central review of imaging
Unknown primary 1 (<1%) 1 (<1%) examinations done at week 12, the rates of objective
Type of disease evolution at inclusion response did not differ significantly between the
Metastasis alone 110 (41%) 118 (44%) two groups: at week 12, 108 (57%) of 190 patients in the
Locoregional relapse alone 94 (35%) 98 (36%) TPEx group had an objective response (eight com­
Metastasis and locoregional relapse 65 (24%) 54 (20%) plete response, 100 partial response) compared with
Previous cancer treatment 106 (59%) of 179 patients in the EXTREME group
No 43 (16%) 53 (20%) (11 complete response, 95 partial response; p=0·64). Best
Yes 226 (84%) 217 (80%) overall response was also not significantly different
Previous platinum agent administration between treatment groups (p=0·89). Few patients
No 113 (42%) 130 (48%) progressed without any previous stabilisation or response
Yes 156 (58%) 140 (52%)
in both groups (table 2).
HPV DNA positivity among patients with oropharyngeal carcinoma†
Details of chemotherapy administration are in table 2
No 84/104 (81%) 62/76 (82%)
and in the appendix (pp 5–6). The median duration of
chemotherapy was 11·1 weeks (IQR 9·0–11·3) in the TPEx
Yes 20/104 (19%) 14/76 (18%)
group and 15·5 weeks (7·6–18·5) in the EXTREME group.
Data are n (%), median (IQR), n/N (%). Percentages might not total 100% due to rounding. ECOG=Eastern Cooperative The proportion of patients receiving the planned number
Oncology Group. EXTREME=cisplatin, fluorouracil, and cetuximab. HPV=human papillomavirus. TPEx=docetaxel,
of cycles was significantly higher in the TPEx group than
cisplatin, and cetuximab. *Multiple initial locations including oropharyngeal site for two patients. †HPV DNA test was
not done in 18 patients with oropharyngeal carcinoma in the TPEx regimen group and in 20 patients with in the EXTREME group (194 [72%] of 269 vs 119 [44%] of
oropharyngeal carcinoma in the EXTREME regimen group. 270; p<0·0001). 92 (10%) of 907 chemotherapy cycles in
the TPEx group were delayed, compared with 308 (27%) of
Table 1: Patient characteristics
1143 in the EXTREME group. Doses of chemotherapy
were less frequently reduced in the TPEx group than in
deaths) among 393 patients randomly assigned until the EXTREME group (appendix pp 5–6). Cisplatin was
Jan 31, 2017: 195 in the TPEx group and 198 in the less frequently replaced by carboplatin in the TPEx group
EXTREME group. Median follow-up was 13·2 months than in the EXTREME group; toxicity-related replacement
(IQR 6·6–18·2) in the TPEx group and 13·9 months of cisplatin with carboplatin was lower in the TPEx group
(7·0–18·4) in the EXTREME group. 82 patients had died in than in the EXTREME group (16 patients [renal toxicities
the TPEx group and 92 patients had died in the EXTREME n=9, ototoxicities n=2, other toxicities n=5] vs 67 patients
group. Median overall survival was 14·8 months (95% CI [renal toxicities n=31, ototoxicities n=10, other toxicities
11·5–17·1) in the TPEx group versus 13·3 months n=26]). More patients received the planned number of
(12·6–15·8) in the EXTREME group. The crude HR for cetuximab administrations during chemotherapy in the

6 www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5

 Articles

TPEx group than in the EXTREME group (137 [51%] of

A
269 vs 82 [30%] of 270; p<0·0001). 100 TPEx: median 14·5 months (95% CI 12·5–15·7)
Among the patients who received chemotherapy, more EXTREME: median 13·4 months (95% CI 12·2–15·4)
90
started maintenance therapy in the TPEx group than the HR 0·89 (95% CI 0·74–1·08); p=0·23
80
EXTREME group (p<0·0001; table 2). The main reasons
patients did not receive maintenance therapy in both 70

Overall survival (%)

groups were adverse events and disease progression 60
(appendix p 6). The median duration of main­tenance was 50
14·1 weeks in both groups (IQR 8·1–30·0 in the TPEx 40
group, 6·1–26·3 in the EXTREME group). 18 patients in 30
the EXTREME group received cetuximab administrations
20
every 2 weeks during maintenance therapy (instead of
10
every week per protocol). Among patients who stopped
maintenance, the main reason for maintenance discon­ 0
0 6 12 18 24 30 36 42 48
tinuation was disease progression in both groups Number at risk
(appendix p 6). (number censored)
The proportion of QOL questionnaires returned was TPEx 269 (0) 215 (2) 155 (2) 96 (6) 64 (20) 42 (32) 24 (42) 12 (49) 3 (57)
EXTREME 270 (0) 205 (3) 150 (3) 95 (8) 51 (21) 27 (38) 17 (42) 7 (46) 2 (50)
similar in the two groups: 247 (92%) of 269 expected at
baseline, 153 (63%) of 242 expected at week 12, 100 (43%) B
of 232 expected at week 18, and 81 (38%) of 215 expected 100 TPEx: median 6·0 months (95% CI 5·7–6·4)
at week 26 in the TPEx group; 241 (89%) of 270 expected 90
EXTREME: median 6·2 months (95% CI 5·8–6·7)
HR 0·88 (95% CI 0·74–1·04); p=0·14
at baseline, 143 (59%) of 244 expected at week 12, 105 (46%) 80
of 227 expected at week 18, and 81 (40%) of 205 expected
Progression-free survival (%)

70
at week 26 in the EXTREME group. The QOL results for
60
all scores of the QLQ-C30 are in the appendix (pp 20–23).
Better quality of life was seen in the TPEx group than in 50

the EXTREME group for global health status, physical 40

functioning, and role functioning, but no significant 30
difference was seen between the groups for the remaining 20
QLQ-C30 scores (appendix p 23). 10
Safety was assessed in 528 patients who received at least
0
one dose of chemotherapy or cetuximab, 263 in the TPEx 0 6 12 18 24 30 36 42 48
group and 265 in the EXTREME group (table 3; appendix Time since randomisation (months)
Number at risk
pp 9–18). 37 patients died in association with adverse (number censored)
events: 16 (eight treatment related) in the TPEx group and TPEx 269 (0) 135 (2) 62 (2) 32 (3) 24 (6) 13 (10) 9 (14) 6 (17) 1 (20)
21 (11 treatment related) in the EXTREME group (for the EXTREME 270 (0) 141 (3) 43 (3) 20 (4) 15 (6) 6 (9) 5 (10) 4 (11) 1 (14)

full list of treatment-related deaths see appendix p 8). Figure 2: Kaplan-Meier estimates of overall survival and progression-free survival
Seven patients in each group had fatal infections (A) Overall survival. (B) Progression-free survival. Point estimates of overall survival and progression-free survival
(including febrile neutropenia). Deaths assessed as at 12, 24, and 36 months with Rothman 95% CIs (vertical bars) are shown.
treatment related were most frequently fatal infections,
six in each group, including four sepsis or septic shock in (60 [23%] vs 100 [38%]), thrombocytopenia (six [2%] vs
each group. Although it was planned that safety would be 52 [20%]), anaemia (22 [8%] vs 53 [20%]), kalaemia
assessed in three categories (no adverse events, highest disorder (25 [10%] vs 60 [23%]), and magnesaemia
grade 1–2, highest grade ≥3), because only one patient disorder (34 [13%] vs 58 [22%]). In the TPEx group, 118
had no adverse events, this patient was analysed with (45%) of 263 patients had at least one serious adverse
patients with grade 1–2 adverse events. Fewer patients event versus 143 (54%) of 265 patients in the EXTREME
had at least one adverse event of grade 3 or worse in the group. We observed more hearing toxicity, of all grades
TPEx group (214 [81%] of 263 patients) than in the combined, in the EXTREME group than in the TPEx
EXTREME group (246 [93%] of 265 patients; p<0·0001). group (table 3). The most common serious adverse events
95 (36%) of 263 patients in the TPEx group had adverse were infections (37 [14%] in the TPEx group vs 41 [15%] in
events of grade 4 or worse, compared with 138 (52%) of the EXTREME group), febrile neutropenia (21 [8%] vs
265 patients in the EXTREME group. The most common ten [4%]), vomiting (seven [3%] vs 20 [8%]), and gen­
grade 3 or worse adverse events were haematological eral physical health deterioration (ten [4%] vs 17 [6%];
events and electrolyte disturbances in both groups. These appendix p 19).
adverse events occurred less frequently in the TPEx group Post-hoc sensitivity analyses excluding the 22 non-
than in the EXTREME group: neutropenia (65 [25%] of eligible patients showed similar results to the main
263 patients vs 130 [49%] of 265 patients), leukopenia analyses for overall survival and progression-free

www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5 7

 Articles

survival (crude HR for death 0·89 [95% CI 0·73–1·08],

TPEx regimen group EXTREME regimen group
(n=269) (n=270) crude HR for progression events or deaths 0·89
[0·74–1·06]).
Number of chemotherapy cycles received
Prespecified exploratory subgroup analyses of overall
0 8 (3%) 6 (2%)
survival and progression-free survival did not show a
1 23 (9%) 32 (12%)
difference between groups according to sex, age,
2 27 (10%) 29 (11%)
tumour location, type of evolution, or HPV status in
3 16 (6%) 25 (9%)
patients with oropharyngeal carcinoma (figure 3).
4 194 (72%) 32 (12%)
However, in patients with an ECOG performance
5 1 (<1%) 27 (10%)
status of 0, there was a greater difference in overall
6 0 (<1%) 119 (44%)
survival and progression-free survival between the
Median 4 (3–4) 5 (3–6) TPEx regimen and the EXTREME regimen than in
Reason for chemotherapy discontinuation*† patients with an ECOG performance status of 1
End of chemotherapy period 191 (73%) 117 (44%) (figure 3). In the EXTREME group, a post-hoc
Adverse event 31 (12%) 57 (22%) exploratory analysis comparing overall survival for
Tumour progression 13 (5%) 35 (13%) patients who received G-CSF support during the
Death 10 (4%) 21 (8%) chemotherapy phase with those who did not suggested
Patient refusal or lost to follow-up 7 (3%) 19 (7%) that G-CSF support could improve overall survival
Other reason 8 (3%) 14 (5%) (pp 28–29). In the EXTREME group, in the post-hoc
Maintenance therapy with cetuximab† analysis of the local investigator evaluation, 109 (40%)
No 72 (28%) 126 (48%) of 270 patients had an objective response at week 12,
Yes 189 (72%) 138 (52%) increasing to a best objective response in 154 (57%)
Best tumour response during treatment patients during therapy. In a post-hoc multivariable
Complete response 25 (9%) 15 (6%) analysis, independent prognostic factors for overall
Partial response 130 (48%) 139 (51%) survival were ECOG performance status and HPV DNA
Stable disease 69 (26%) 62 (23%) status (appendix pp 24–25). Independent prognostic
Progressive disease 21 (8%) 29 (11%) factors for progression-free survival were age, ECOG
Not evaluable or not evaluated 24 (9%) 25 (9%) performance status, and type of evolution (appendix
pp 24–25).
Data are n (%) or median (IQR). Percentages might not total 100% due to rounding. EXTREME=cisplatin,
fluorouracil, and cetuximab. TPEx=docetaxel, cisplatin, and cetuximab. *Reason unknown for one patient in each
501 patients were evaluable for post-hoc analysis of
group. †Reason for discontinuation and starting of maintenance therapy shown only in patients who started second-line treatment: 245 in the TPEx group and 256 in
chemotherapy (TPEX n=261, EXTREME n=264). the EXTREME group. 157 (64%) of 245 patients in the
Table 2: Treatment and tumour response
TPEx group received second-line treatment versus
164 (64%) of 256 patients in the EXTREME group. Overall

TPEx regimen group (n=263) EXTREME regimen group (n=265)

Any grade Grade 1–2 Grade 3 Grade 4 Grade 5 Any grade Grade 1–2 Grade 3 Grade 4 Grade 5
Any type of adverse event† 263 (100%) 260 (99%) 193 (73%) 87 (33%) 16 (6%) 264 (100%) 258 (97%) 231 (87%) 123 (46%) 21 (8%)
Blood system disorders
Anaemia 204 (78%) 182 (69%) 19 (7%) 2 (1%) 1 (<1%) 216 (82%) 163 (62%) 51 (19%) 2 (1%) 0
Neutropenia 110 (42%) 45 (17%) 22 (8%) 43 (16%) 0 181 (68%) 51 (19%) 70 (26%) 60 (23%) 0
Leukopenia 118 (45%) 58 (22%) 38 (14%) 22 (8%) 0 171 (65%) 71 (27%) 77 (29%) 23 (9%) 0
Thrombocytopenia 98 (37%) 92 (35%) 4 (2%) 1 (<1%) 1 (<1%) 166 (63%) 114 (43%) 34 (13%) 18 (7%) 0
Febrile neutropenia 26 (10%) 2 (1%) 13 (5%) 8 (3%) 3 (1%) 16 (6%) 1 (<1%) 7 (3%) 8 (3%) 0
Metabolism and nutrition disorders
Magnesaemia disorder 156 (59%) 122 (46%) 25 (10%) 9 (3%) 0 178 (67%) 120 (45%) 41 (15%) 17 (6%) 0
Kalaemia disorder 135 (51%) 110 (42%) 19 (7%) 6 (2%) 0 176 (66%) 116 (44%) 46 (17%) 14 (5%) 0
Calcaemia disorder 153 (58%) 138 (52%) 7 (3%) 8 (3%) 0 148 (56%) 132 (50%) 12 (5%) 4 (2%) 0
Natraemia disorder 127 (48%) 108 (41%) 17 (6%) 2 (1%) 0 142 (54%) 111 (42%) 28 (11%) 3 (1%) 0
Phosphataemia disorder 128 (49%) 108 (41%) 19 (7%) 1 (<1%) 0 140 (53%) 112 (42%) 25 (9%) 3 (1%) 0
Hyperglycaemia 48 (18%) 46 (17%) 2 (1%) 0 0 64 (24%) 52 (20%) 9 (3%) 3 (1%) 0
Anorexia 86 (33%) 73 (28%) 13 (5%) 0 0 85 (32%) 70 (26%) 14 (5%) 1 (<1%) 0
Weight loss 52 (20%) 50 (19%) 2 (1%) 0 0 58 (22%) 54 (20%) 4 (2%) 0 0
(Table 3 continues on next page)

8 www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5

 Articles

TPEx regimen group (n=263) EXTREME regimen group (n=265)

Any grade Grade 1–2 Grade 3 Grade 4 Grade 5 Any grade Grade 1–2 Grade 3 Grade 4 Grade 5
(Continued from previous page)
Investigations
Gamma-glutamyltransferase increased 115 (44%) 104 (40%) 11 (4%) 0 0 139 (52%) 123 (46%) 12 (5%) 4 (2%) 0
Serum albumin decreased 137 (52%) 132 (50%) 5 (2%) 0 0 133 (50%) 125 (47%) 8 (3%) 0 0
Alkaline phosphatase increased 64 (24%) 64 (24%) 0 0 0 66 (25%) 66 (25%) 0 0 0
Alanine aminotransferase increased 58 (22%) 57 (22%) 1 (<1%) 0 0 58 (22%) 57 (22%) 0 1 (<1%) 0
Aspartate aminotransferase increased 56 (21%) 55 (21%) 1 (<1%) 0 0 54 (20%) 51 (19%) 1 (<1%) 2 (1%) 0
Creatinine increased 69 (26%) 60 (23%) 4 (2%) 4 (2%) 1 (<1%) 119 (45%) 98 (37%) 17 (6%) 4 (2%) 0
General disorders
Fatigue 179 (68%) 144 (55%) 34 (13%) 1 (<1%) 0 195 (74%) 147 (55%) 47 (18%) 1 (<1%) 0
Fever 31 (12%) 25 (10%) 5 (2%) 0 1 (<1%) 32 (12%) 29 (11%) 2 (1%) 1 (<1%) 0
General physical health deterioration 12 (5%) 4 (2%) 7 (3%) 0 1 (<1%) 21 (8%) 5 (2%) 14 (5%) 2 (1%) 0
Allergic reaction 12 (5%) 6 (2%) 3 (1%) 3 (1%) 0 20 (8%) 7 (3%) 7 (3%) 5 (2%) 1 (<1%)
Skin and subcutaneous tissue disorders
Rash acneiform 168 (64%) 156 (59%) 11 (4%) 1 (<1%) 0 167 (63%) 150 (57%) 16 (6%) 1 (<1%) 0
Dry skin 74 (28%) 74 (28%) 0 0 0 72 (27%) 71 (27%) 1 (<1%) 0 0
Palmar–plantar erythrodysesthesia 28 (11%) 25 (10%) 3 (1%) 0 0 36 (14%) 32 (12%) 4 (2%) 0 0
syndrome
Alopecia 60 (23%) 58 (22%) 1 (<1%) 1 (<1%) 0 31 (12%) 31 (12%) 0 0 0
Gastrointestinal disorders
Mucositis oral 122 (46%) 102 (39%) 16 (6%) 4 (2%) 0 153 (58%) 118 (45%) 34 (13%) 1 (<1%) 0
Nausea 135 (51%) 125 (48%) 10 (4%) 0 0 173 (65%) 145 (55%) 28 (11%) 0 0
Vomiting 84 (32%) 74 (28%) 9 (3%) 1 (<1%) 0 116 (44%) 87 (33%) 29 (11%) 0 0
Diarrhoea 116 (44%) 102 (39%) 14 (5%) 0 0 94 (35%) 77 (29%) 16 (6%) 1 (<1%) 0
Constipation 61 (23%) 61 (23%) 0 0 0 76 (29%) 75 (28%) 1 (<1%) 0 0
Dysphagia 33 (13%) 24 (9%) 9 (3%) 0 0 48 (18%) 32 (12%) 16 (6%) 0 0
Infection (any type) 90 (34%) 48 (18%) 29 (11%) 8 (3%) 5 (2%) 89 (34%) 43 (16%) 29 (11%) 10 (4%) 7 (3%)
Ear disorders
Tinnitus 14 (5%) 13 (5%) 1 (<1%) 0 0 33 (12%) 30 (11%) 3 (1%) 0 0
Hearing impairment or hypoacusis 10 (4%) 9 (3%) 1 (<1%) 0 0 30 (11%) 24 (9%) 6 (2%) 0 0
Other events
Hypotension 22 (8%) 17 (6%) 5 (2%) 0 0 25 (9%) 19 (7%) 6 (2%) 0 0
Peripheral sensory neuropathy 29 (11%) 28 (11%) 1 (<1%) 0 0 26 (10%) 25 (9%) 1 (<1%) 0 0
Dyspnoea 25 (10%) 17 (6%) 4 (2%) 2 (1%) 2 (1%) 32 (12%) 28 (11%) 2 (1%) 1 (<1%) 1 (<1%)
Data are n (%). Percentages might not total 100% due to rounding. EXTREME=cisplatin, fluorouracil, and cetuximab. TPEx=docetaxel, cisplatin, and cetuximab. *Includes grade 1 or 2 adverse events occurring in
≥10% of patients in either group, and grade 3, 4, or 5 adverse events occurring in ≥2% patients in either group. A complete list of grade 3–5 adverse events is in the appendix (pp 9–18). †Patients who had
different adverse events of different grades are counted in each grade for which they had at least one adverse event; therefore, the number of patients with adverse events of any grade is not the sum of patients
with adverse events of grades 1–2, 3, 4, and 5.

Table 3: Adverse events during the chemotherapy phase*

survival results were encouraging in patients who compared with the median overall survival of 10·1 months
received TPEx or EXTREME followed by immunotherapy. with the EXTREME regimen in the trial by Vermorken
Detailed results of the post-hoc analysis of second-line and colleagues.1 There is no clear reason for such a
treatment are in the appendix (pp 26–27). variation in median overall survival beyond possible
gradual improvements of outcomes due to improved
Discussion supportive care since the previous study in 2008, but
Our study did not show a benefit of TPEx compared with several differences between the two studies can be
EXTREME in terms of overall survival in patients with emphasised. First, initial treatments received before trial
recurrent or metastatic HNSCC, and the primary objective enrolment differed between the two studies: 41% of
was not met. However, median overall survival of patients patients in Vermorken and colleagues’ study1 were exposed
in the TPEx group was long (14·5 months), in line with to a previous platinum drug, compared with 52% in our
our previous phase 2 trial.19 In the EXTREME group, EXTREME regimen group. Improved imaging technology
median overall survival was also long (13·4 months), in the past few years (CT and PET scans) could have led to

www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5 9

 Articles

A
TPEx EXTREME Hazard ratio p value
(95% CI)

Deaths/ Median overall survival, Deaths/ Median overall survival,

patients months (95% CI) patients months (95% CI)

Sex
Male 189/240 14·1 (12·2–15·4) 190/231 13·4 (11·5–15·8) 0·91 (0·74–1·11) 0·53
Female 20/29 19·9 (11·0–32·7) 28/39 13·6 (8·5–20·4) 0·73 (0·41–1·32)
Age, years
<60 94/127 14·1 (11·4–15·9) 109/133 12·9 (11·1–15·4) 0·81 (0·62–1·07) 0·39
≥60 115/142 14·8 (12·4–17·2) 109/137 15·0 (11·3–17·3) 0·97 (0·74–1·26)
ECOG performance status
0 56/86 20·9 (14·9–33·5) 67/86 17·6 (15·0–21·7) 0·66 (0·46–0·95) 0·072
1 153/183 12·5 (9·8–14·8) 151/184 11·3 (9·2–13·6) 1·01 (0·81–1·27)
Tumour location
Oropharynx 89/123 15·2 (12·5–17·6) 76/96 12·5 (9·6–15·8) 0·76 (0·56–1·04) 0·28
Larynx or hypopharynx 71/88 14·1 (11·5–17·4) 97/120 15·0 (12·9–17·8) 1·02 (0·75–1·39)
Oral cavity or other 49/58 11·3 (9·1–15·2) 45/54 12·6 (8·5–17·6) 1·06 (0·71–1·60)
Type of evolution
Metastasis alone 80/110 14·5 (11·3–18·7) 93/118 16·4 (12·8–18·2) 0·87 (0·65–1·18) 0·91
Metastasis and locoregional relapse 53/65 11·5 (8·0–15·2) 42/54 9·6 (6·3–15·8) 0·95 (0·63–1·43)
Locoregional relapse alone 76/94 15·1 (12·6–16·9) 83/98 12·9 (9·3–15·4) 0·84 (0·62–1·15)
HPV status in patients with oropharyngael carcinoma
HPV DNA negative 64/84 14·5 (11·1–16·9) 53/62 11·3 (8·6–15·2) 0·72 (0·50–1·04) 0·99
HPV DNA positive 10/20 36·2 (15·2–NR) 8/14 21·1 (9·6–NR) 0·74 (0·29–1·88)
Overall 209/269 14·5 (12·5–15·7) 218/270 13·4 (12·2–15·4) 0·89 (0·74–1·08)

B
TPEx EXTREME Hazard ratio p value
(95% CI)

Events/ Median progression– Events/ Median progression–

patients free survival, months patients free survival, months
(95% CI) (95% CI)

Sex
Male 221/240 6·0 (5·6–6·4) 220/231 6·1 (5·8–6·6) 0·86 (0·72–1·04) 0·72
Female 27/29 6·3 (4·7–8·8) 35/39 7·3 (5·6–8·0) 0·97 (0·59–1·61)
Age, years
<60 113/127 6·1 (5·4–7·7) 124/133 6·2 (5·6–6·8) 0·82 (0·64–1·06) 0·38
≥60 135/142 5·8 (5·1–6·4) 131/137 6·5 (5·6–7·1) 0·93 (0·73–1·19)
ECOG performance status
0 74/86 8·1 (6·1–9·7) 81/86 6·6 (5·8–7·7) 0·65 (0·47–0·90) 0·021
1 174/183 5·5 (4·7–6·0) 174/184 6·0 (5·1–6·9) 1·03 (0·83–1·27)
Tumour location
Oropharynx 110/123 6·0 (5·3–7·1) 90/96 6·1 (5·1–6·9) 0·83 (0·63–1·10) 0·71
Larynx or hypopharynx 83/88 6·1 (4·9–7·9) 114/120 6·4 (5·7–7·2) 0·89 (0·67–1·19)
Oral cavity or other 55/58 5·9 (5·1–8·0) 51/54 6·4 (4·7–7·9) 0·97 (0·66–1·42)
Type of evolution
Metastasis alone 103/110 5·8 (5·0–6·1) 114/118 6·6 (6·0–7·1) 0·99 (0·76–1·30) 0·57
Metastasis and locoregional relapse 61/65 5·6 (4·3–6·5) 49/54 5·1 (4·0–6·5) 0·83 (0·57–1·21)
Locoregional relapse alone 84/94 7·8 (6·0–8·6) 92/98 6·2 (4·8–7·8) 0·81 (0·60–1·09)
HPV status in patients with oropharyngael carcinoma
HPV DNA negative 75/84 6·2 (5·5–7·9) 58/62 6·1 (4·7–7·7) 0·84 (0·59–1·18) 0·91
HPV DNA positive 17/20 6·3 (4·5–21·6) 12/14 6·2 (4·4–12·7) 0·87 (0·41–1·84)
Overall 248/269 6·0 (5·7–6·4) 255/270 6·2 (5·8–6·7) 0·88 (0·74–1·04)

0·2 1·0 2·0

Favours Favours
TPEx EXTREME

10 www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5

 Articles

earlier detection of relapse or metastases in our study G-CSF support, the substitution of fluorouracil by
(47% of patients presented with metastases in Vermorken docetaxel, and a lower dose of cisplatin. However, the
and colleagues’ study1 vs 64% in our EXTREME group). TPEx regimen is still a chemotherapy-based regimen
Second, patients with an ECOG performance status of 2 with clinically significant toxicities and should be reserved
were excluded from our study, whereas they represented for fit patients in whom a rapid tumour response is
12% of patients in Vermorken and colleagues’ study.1 needed. Consistent with the toxicity results and with the
Third, G-CSF administration, which was mandatory in shorter duration of chemotherapy, analysis of QOL also
the TPEx group in our study, was also used in suggested benefits in some scales on the QLQ-C30 (global
43% of patients (33% of the cycles) in the EXTREME health status, physical functioning, and role functioning)
group. Our study was done in countries with widespread in favour of TPEx compared with EXTREME, with the
experience with the EXTREME regimen, and perhaps other scales showing no significant differences between
more routine use of G-CSF in this setting. An exploratory the groups.
analysis suggested an increased overall survival for Other first-line taxane-based combinations such as
patients who received G-CSF support during the paclitaxel–carboplatin with cetuximab have been tested in
chemotherapy phase, compared with those who did not. several studies but have not been compared against the
This outcome could be related to a decreased toxicity of standard of care in large randomised trials as has been
the chemotherapy–cetuximab combination, and in favour done in our trial. These studies also showed promising
of a systematic use of G-CSF support. In agreement with safety and efficacy results for fit or unfit patients.21
the overall survival analysis, progression-free survival and PD-1 inhibition has previously shown promising
objective response rates did not differ between the two results22 in the second-line treatment of recurrent or
groups. In both groups, the proportion of patients who metastatic HNSCC. In 2019, the Keynote 048 randomised
had disease progression without any previous stabilisation phase 3 trial8 compared the EXTREME regimen with
or response was very low (8% in the TPEx group vs 11% in pembrolizumab alone or pembrolizumab combined with
the EXTREME group). However, the substitution of platinum–fluorouracil and reported significant overall
fluorouracil by docetaxel in the TPEx regimen led to survival improvements in both pembrolizumab groups
interesting findings regarding compliance and toxicity. for patients with a combined positive score for PD-L1
Treatment compliance was better in the TPEx group than expression of 1 or more compared with the EXTREME
the EXTREME group, with significantly more delays in group (median overall survival in the patient population
chemotherapy, more dose adjustments, and more with a combined positive score for PD-L1 expression of 1
frequent switch to carbo­platin in the EXTREME group. or more: pembrolizumab alone 12·3 months; cisplatin–
Carboplatin switches could cause an excess of fluorouracil–pembrolizumab 13·6 months; EXTREME
haematological toxicity in this group, even higher than 10·4 months), although there were no significant
that found by Vermorken and colleagues.1 The fact that all differences in progression-free survival across the
patients in our study started with cisplatin might have three groups.8–10 Consequently, the EXTREME regimen
affected the safety profile by causing a higher electrolyte was replaced in 2020 by this new standard of care of
toxicity than in the study by Vermorken and colleagues. pembrolizumab alone or combined with platinum and
The better compliance in the TPEx group compared with fluorouracil for patients with a PD-L1 combined positive
the EXTREME group was observed even during the first score of 1 or more.
four cycles of treatment. The TPEx regimen was shorter Our study has some limitations. The proportion of
and the proportion of patients who received the planned patients in our population who were HPV DNA positive
treatment was significantly higher than in the EXTREME was low, and possibly somewhat underestimated by the
group, with a significantly higher proportion of patients HPV DNA test used. However, the study was done in
entering the maintenance phase. Another advantage of countries where most patients with HNSCC are current
the TPEx regimen, other than limiting the chemotherapy or former smokers, with a lower HPV incidence
at four cycles, is the cetuximab maintenance every 2 weeks compared with the USA or Scandinavia. The trial was
instead of weekly, which reduces patient constraints designed for efficacy, not for non-inferiority because of
without jeopardising efficacy. the very promising overall survival results obtained in
TPEx was also substantially less toxic than EXTREME the initial phase 2 study (median overall survival of
and patients in the TPEx group had fewer adverse events 14·0 months19) compared with the median overall survival
of grade 3 or worse, which might be due in part to more of 10·1 months with the EXTREME regimen at that time.
However, overall survival in the EXTREME group in our
trial was much longer than previously reported and
expected. Long-term data on QOL are not available,
Figure 3: Subgroup analysis of overall survival and progression-free survival because QOL questionnaires were only com­pleted until
(A) Overall survival. (B) Progression-free survival. The area of each square is
proportional to the number of events in the subgroup. ECOG=Eastern Cooperative
week 26, but this fully covered the period of chemo­
Oncology Group. EXTREME=cisplatin, fluorouracil, cetuximab. HPV=human therapy as well as the early maintenance treatment
papillomavirus. NR=not reached. TPEx=docetaxel, cisplatin, cetuximab. phase. The study was started before the introduction of

www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5 11

 Articles

immunotherapy for HNSCC and the EXTREME regimen BMS, Innate Pharma, and Merck, all outside the submitted work. AA has
is no longer the standard of care for all patients in this received grant support, paid to her institution, from the Groupe
d’Oncologie Radiotherapie Tête et Cou (GORTEC; French Radiation and
setting. However, pembrolizumab alone or combined Oncology Group for Head and Neck) for this study and has received grant
with platinum–fluorouracil is not available for all patients support, paid to her institution, from F Hoffmann–La Roche, outside the
with recurrent or metastatic HNSCC and is not beneficial submitted work. JF has received grants, personal fees, and non-financial
to all patients. The European Medicines Agency did not support from AstraZeneca and BMS, personal fees and non-financial
support from Merck Sharp & Dohme (MSD), and personal fees from
approve pembrolizumab alone or added to platinum– Merck and Innate, outside the submitted work. AF has received grant
fluorouracil for patients with a negative PD-L1 combined support, paid to his institution, from the GORTEC for this study. UK has
positive score.9 received grants and personal fees from Merck, outside the submitted
Based on all these data, the TPEx regimen might offer work. JB participated in advisory boards for MSD, BMS, Debiopharm,
Merck, and AstraZeneca, outside the submitted work. RM has received
an alternative to the EXTREME regimen in first-line research grants from Merck-Serono and has been an advisory board
treatment of many patients with recurrent or metastatic member for AstraZeneca, BMS, Rakuten, Merck-Serono, MSD,
HNSCC, especially for those with a negative PD-L1 Nanobiotics, Bayer, and Roche, all outside the submitted work. PS has
combined positive score, those who might not be good received personal fees from Merck Serono, outside the submitted work.
CEve has received personal fees from AstraZeneca, BMS, Innate Pharma,
candidates for up-front pembrolizumab due to immuno­ MSD, and Merck Serono, all outside the submitted work. LG has received
logically relevant comorbidities, patients with a high personal fees from BMS, IPSEN, Merck Serono, MSD, and Pfizer,
tumour burden or symptoms that mean a rapid response outside the submitted work. BL has received personal fees from
is a key treatment goal,10 or for patients with contra­ AstraZeneca, BMS, IPSEN, Janssen, MSD, and Roche, outside the
submitted work. All other authors declare no competing interests.
indication to fluorouracil. The post-hoc analysis on
second-line therapies showed good overall survival results Data sharing
Individual participant data and other data and documents will not be
for patients in the TPEx group who received second-line shared.
immunotherapy with PD-1 inhibitors or PD-L1 inhibitors
Acknowledgments
(median overall survival 21·9 months). Although this was The trial was sponsored by the GORTEC and was done in collaboration
a post-hoc analysis, this is an interesting finding and with the Tratamiento de Tumores de Cabeza y Cuello (TTCC; Spanish
suggests that a treatment sequence of TPEx followed Head and Neck Cancer Cooperative Group) and the Arbeitsgemeinschaft
Internistische Onkologie (AIO; German Association of Medical
by anti-PD-1 or anti-PD-L1 requires further testing.
Oncology group). We thank the patients and their families, the
Alternatively, inhibitors of PD-1 or PD-L1 could be GORTEC, TTCC, and AIO investigators and colleagues at the many
introduced earlier into the TPEx regimen in association centres in this trial, the Independent Data Safety Monitoring Committee
with cetuximab in the maintenance phase. members (Véronique Mosseri [Institut Curie, Paris, France],
Marco Merlano [S Croce & Carle University teaching Hospital, Cuneo,
In conclusion, this randomised trial confirmed the good
Italy], and Jean-Pascal Machiels [Cliniques Universitaires Sain-Luc,
survival results of the TPEx regimen previously observed Brussels, Belgium]), the members of the independent imaging review
in the initial phase 2 trial. Compared with the EXTREME committee (François Bidault, Sami Ammari [Gustave Roussy, Villejuif,
regimen, TPEx did not show a significant benefit in overall France], Sophie Espinoza [Paris, France], and Arthur Varoquaux [APHM,
Marseille, France]), the study team at GORTEC (M Bindzi,
survival, but was a shorter and better-tolerated treatment M Delhommeau, M Khalid, A Pechery, C Rauche, C Michel,
regimen, and showed a better QOL. The TPEx regimen N Vintonenko, and Raissa Kapso [Biostatistic and Epidemiology Unit,
might offer an alternative to the EXTREME regimen or Gustave Roussy, Villejuif ]). Merck Santé (Lyon, France), an affiliate of
pembrolizumab in first-line treatment of fit patients with Merck (Darmstadt, Germany) supported the trial by providing a research
grant and docetaxel at no cost. Chugai Pharma (Paris La Défense,
recurrent or metastatic HNSCC. France) also provided a research grant.
Contributors
References
JG, AA, and JB contributed to the conception and design of the study 1 Vermorken JB, Mesia R, Rivera F, et al. Platinum-based
and statistical analysis plan, and have accessed and verified the data. chemotherapy plus cetuximab in head and neck cancer.
AA did the statistical analyses. CC-C assembled the data and has N Engl J Med 2008; 359: 1116–27.
accessed and verified the data. AF has accessed and verified the data 2 Grégoire V, Lefebvre JL, Licitra L, Felip E. Squamous cell carcinoma
and took part in the statistical analyses. JG, JF, ES-B, CL, MT, LG, LM, of the head and neck: EHNS-ESMO-ESTRO Clinical Practice
OC, DC, HC, DV, PS, AJ, CEve, CS, SD, CEvr, J-PD, BL, SZ, UK, JB, Guidelines for diagnosis, treatment and follow-up. Ann Oncol 2010;
and RM treated patients and collected data. FL contributed to the 21 (suppl 5): v184–86.
writing of the manuscript. LS managed the project. All authors 3 Peyrade F, Cupissol D, Geoffrois L, et al. Systemic treatment and
contributed to the data analysis and interpretation. All authors had full medical management of metastatic squamous cell carcinoma of the
access to all of the data in the study and take responsibility for the head and neck: review of the literature and proposal for
integrity of the data and the accuracy of the data analysis. All authors management changes. Oral Oncol 2013; 49: 482–91.
contributed to the drafting of the manuscript or critical revision of the 4 Argiris A, Kotsakis AP, Hoang T, et al. Cetuximab and bevacizumab:
manuscript for important intellectual content. All authors provided preclinical data and phase II trial in recurrent or metastatic squamous
final approval of the manuscript, and had final responsibility for the cell carcinoma of the head and neck. Ann Oncol 2013; 24: 220–25.
decision to submit for publication. All authors agree to be accountable 5 Vermorken JB, Licitra L, Stöhlmacher-Williams J, et al. Phase II
for all aspects of the work in ensuring that questions related to the study of pemetrexed in combination with cisplatin and cetuximab
accuracy or integrity of any part of the work are appropriately in recurrent or metastatic squamous cell carcinoma of the head and
neck. Eur J Cancer 2013; 49: 2877–83.
investigated and resolved.
6 Vermorken JB, Peyrade F, Krauss J, et al. Cisplatin, 5-fluorouracil,
Declaration of interests and cetuximab (PFE) with or without cilengitide in recurrent/
JG has received research grants from Boehringer Ingelheim, Bristol metastatic squamous cell carcinoma of the head and neck: results
Myers Squibb (BMS), Chugai, GlaxoSmithKline (GSK), Merck Serono, of the randomized phase I/II ADVANTAGE trial (phase II part).
and Sanofi and has been an advisory board member for AstraZeneca, Ann Oncol 2014; 25: 682–88.

12 www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5

 Articles

7 Vermorken JB, Stöhlmacher-Williams J, Davidenko I, et al. 16 Kies MS, Holsinger FC, Lee JJ, et al. Induction chemotherapy and
Cisplatin and fluorouracil with or without panitumumab in patients cetuximab for locally advanced squamous cell carcinoma of the
with recurrent or metastatic squamous-cell carcinoma of the head head and neck: results from a phase II prospective trial. J Clin Oncol
and neck (SPECTRUM): an open-label phase 3 randomised trial. 2010; 28: 8–14.
Lancet Oncol 2013; 14: 697–710. 17 Specht L, Larsen SK, Hansen HS. Phase II study of docetaxel and
8 Burtness B, Harrington KJ, Greil R, et al. Pembrolizumab alone or cisplatin in patients with recurrent or disseminated squamous-cell
with chemotherapy versus cetuximab with chemotherapy for carcinoma of the head and neck. Ann Oncol 2000; 11: 845–49.
recurrent or metastatic squamous cell carcinoma of the head and 18 Knoedler M, Gauler TC, Gruenwald V, et al. Phase II study of
neck (KEYNOTE-048): a randomised, open-label, phase 3 study. cetuximab in combination with docetaxel in patients with recurrent
Lancet 2019; 394: 1915–28. and/or metastatic squamous cell carcinoma of the head and neck
9 European Medicines Agency. Assessment report EMA/ after platinum-containing therapy: a multicenter study of the
CHMP/591139/2019. https://www.ema.europa.eu/en/documents/ Arbeitsgemeinschaft Internistische Onkologie. Oncology 2013;
variation-report/keytruda-h-c-3820-ii-0065-epar-assessment-report- 84: 284–89.
variation_en.pdf (accessed Jan 19, 2020). 19 Guigay J, Fayette J, Dillies AF, et al. Cetuximab, docetaxel, and
10 Szturz P, Vermorken JB. Translating KEYNOTE-048 into practice cisplatin as first-line treatment in patients with recurrent or
recommendations for head and neck cancer. Ann Transl Med 2020; metastatic head and neck squamous cell carcinoma: a multicenter,
8: 975. phase II GORTEC study. Ann Oncol 2015; 26: 1941–47.
11 Rose WC, Wild R. Therapeutic synergy of oral taxane BMS-275183 20 Klinghammer K, Gauler T, Dietz A, et al. Cetuximab, fluorouracil
and cetuximab versus human tumor xenografts. Clin Cancer Res and cisplatin with or without docetaxel for patients with recurrent
2004; 10: 7413–17. and/or metastatic squamous cell carcinoma of the head and neck
12 Argiris A, Heron DE, Smith RP, et al. Induction docetaxel, cisplatin, (CeFCiD): an open-label phase II randomised trial
and cetuximab followed by concurrent radiotherapy, cisplatin, (AIO/IAG-KHT trial 1108). Eur J Cancer 2019; 122: 53–60.
and cetuximab and maintenance cetuximab in patients with locally 21 Guigay J, Tahara M, Licitra L, et al. The evolving role of taxanes in
advanced head and neck cancer. J Clin Oncol 2010; 28: 5294–300. combination with cetuximab for the treatment of recurrent and/or
13 Gedlicka C, Formanek M, Selzer E, et al. Phase II study with metastatic squamous cell carcinoma of the head and neck: evidence,
docetaxel and cisplatin in the treatment of recurrent and/or advantages, and future directions. Front Oncol 2019; 9: 668.
metastatic squamous cell carcinoma of the head and neck. Oncology 22 Ferris RL, Blumenschein G Jr, Fayette J, et al. Nivolumab vs
2002; 63: 145–50. investigator’s choice in recurrent or metastatic squamous cell
14 Glisson BS, Murphy BA, Frenette G, Khuri FR, Forastiere AA. carcinoma of the head and neck: 2-year long-term survival update of
Phase II trial of docetaxel and cisplatin combination chemotherapy CheckMate 141 with analyses by tumor PD-L1 expression.
in patients with squamous cell carcinoma of the head and neck. Oral Oncol 2018; 81: 45–51.
J Clin Oncol 2002; 20: 1593–99.
15 Hitt R, Irigoyen A, Cortes-Funes H, Grau JJ, García-Sáenz JA,
Cruz-Hernandez JJ. Phase II study of the combination of cetuximab
and weekly paclitaxel in the first-line treatment of patients with
recurrent and/or metastatic squamous cell carcinoma of head and
neck. Ann Oncol 2012; 23: 1016–22.

www.thelancet.com/oncology Published online March 5, 2021 https://doi.org/10.1016/S1470-2045(20)30755-5 13