Chemotherapy in non-small-cell lung cancer: an update of an individual patient data meta-analysis.

Full text

Page 1

and the majority of tissue samples used in current clinical
practice today are archived specimens from numerous differ-
ent hospitals and laboratories, it is impossible to determine
whatfixativeswereused.Inaddition,theseareusuallysamples
of the primary tumor, which may differ in terms of EGFR
expressionfromthemetastasesthatarebeingtreated.
A further important observation of Atkins et al is that
immunoreactivity of anti EGFR is inversely correlated with
storage time, with a substantial drop in the number of 3?
EGFR readings and a substantial increase in the number of
EGFR (?), or zero, readings with an increase in storage
timefrom3to24months,andeventhisdegreeofvariation
varied considerably depending on which fixative had been
used. EGFR analyses in our study and in practice today are
performed on whatever tissue sample is available in ar-
chived specimens. They may have been obtained and pro-
cessedatanynumberofdifferenthospitalsandlaboratories,
fromafewdaysagotomanyyearsago.Thusthefindingsof
EGFR IHC in our study, other published studies, and stan-
dard practice, are likely to be far more random and far less
meaningful than any of us have cared to admit.
Other observations about the nature of EGFR cast fur-
therdoubtsabouttherelevanceofcurrentEGFRIHCtech-
niques.QuantitativeEGF-bindingexperimentshaveshown
thattheEGFRexistsonthecellsurfaceinbothhigh-affinity
andlow-affinityforms.2,3Thehigh-affinity(KD?10to100
pM)untetheredform,whichmayrepresentfrom2%to5%
of total surface receptors, comprises the minority of
the EGFRs quantitatively. The low-affinity (KD? 2 to 5
nM/L) tethered form is largely autoinhibited and accounts
forthemajorityofEGFRsonthecellsurface,quantitatively,
but may account for relatively little of the receptor tyrosine
kinaseactivationnecessaryforinitiationofsignaltransduc-
tion. The commonly used antibodies for EGFR determina-
tion do not discriminate between the high- and low-affinity
receptors.Itisthereforeareasonablehypothesisthatevenifwe
were to overcome the inconsistencies in tissues fixation, stor-
age, and scoring techniques that now plague us, we may not
haveanadequatemarkerforpredictionofclinicalactivity.
Theinescapableconclusionisthatourcurrentattempt
to use EGFR IHC for patient selection for EGFR antagonist
therapy has thus far been a total failure. We should ac-
knowledge this and move forward. Use of currently avail-
able routine IHC measurement of EGFR status should not
be regarded as part of routine patient management, as the
assaysavailablehavenodemonstratedpredictivevalue.Itis
incorrecttoselectapatientoratumortypefortreatmentwith
cetuximab on the basis of high EGFR IHC expression. It is
equally illogical and wrong to withhold cetuximab solely be-
causeatumorisnegativeforEFGRbyIHCstaining.
Leonard Saltz
Memorial Sloan-Kettering Cancer Center, New York, NY
■ ■ ■
Author’s Disclosures of Potential
Conflicts of Interest
The following author or their immediate family mem-
bershaveindicatedafinancialinterest.Noconflictexistsfor
drugs or devices used in a study if they are not being evalu-
ated as part of the investigation. Honoraria: Leonard Saltz,
Bristol Myers Squibb. For a detailed description of these
categories, or for more information about ASCO’s conflict
of interest policy, please refer to the Author Disclosure
Declaration and the Disclosures of Potential Conflicts of
Interest section of Information for Contributors found in
the front of every issue.
REFERENCES
1. Atkins D, Reiffen KA, Tegtmeier CL, et al: Immunohistochemical
detection of EGFR in paraffin-embedded tumor tissues: Variation in staining
intensity due to choice of fixative and storage time of tissue sections.
J Histochem Cytochem 52:893-901, 2004
2. Lax I, Bellot F, Howk R, et al: Functional analysis of the ligand binding
site of EGF-receptor utilizing chimeric chicken/human receptor molecules.
Embo J 8:421-427, 1989
3. Mattoon D, Klein P, Lemmon MA: The tethered configuration of the
EGF receptor extracellular domain exerts only a limited control of receptor
function. Proc Natl Acad Sci U S A 101:923-928, 2004
DOI: 10.1200/JCO.2005.05.354
ChemotherapyinNon–Small-Cell
LungCancer:AnUpdateofan
IndividualPatientDataMeta-Analysis
TO THE EDITOR: Hotta et al1have reported a meta-
analysis of postoperative chemotherapy versus surgery
alone in non–small-cell lung cancer (NSCLC). This meta-
analysis is based on abstracted data from 11 randomized
controlledtrials(RCTs)thatincluded5,716patients.These
11 RCTs have been published since 1995, when an individ-
ual patient data–based (IPD) meta-analysis2assessing the
value of chemotherapy in this disease was published.
The NSCLC Collaborative Group was responsible for
the IPD meta-analysis published in 1995, and we welcome
the meta-analysis provided by Hotta et al,1which was a
thoughtful and well done meta-analysis with data ab-
stracted from publications. However, we wholeheartedly
agree with the Editorial by Piedbois and Buyse3who point
out the inherent difficulties of meta-analyses that use data
from published reports and highlight the advantages and
increased reliability of the IPD approach.
Both publications1,3conclude that an IPD meta-
analysisshouldbecarriedoutinordertoverifytheseresults
and we are pleased to confirm that we are in the process of
updating the 1995 meta-analysis.
The 1995 meta-analysis included more than 9,000 pa-
tients from 52 RCTs. It assessed the effect of chemotherapy
in four different settings. Since the meta-analysis publica-
tion, there has been renewed enthusiasm for investigations
Correspondence
924
JOURNAL OF CLINICAL ONCOLOGY
Downloaded from jco.ascopubs.org on October 14, 2011. For personal use only. No other uses without permission.
Copyright © 2005 American Society of Clinical Oncology. All rights reserved.

Page 2

of chemotherapy in NSCLC, and a considerable number of
new RCTs have been completed. The total number of
patients randomly assigned has risen to approximately
23,000 patients.
As the aim of the NSCLC Collaborative Group is to
provide an up-to-date and reliable review of the role of
chemotherapy, both to act as a sound basis for evidence
based medicine and to help guide future research, it was
decidedthatanupdatewastimely.Anumberofnewagents
and timings have been investigated in all settings, and the
update consists of adding trials published since the 1995
analyses and additional follow-up data from trials already
included, as well as looking at additional outcomes in cer-
tain settings. We are also investigating the effect of chemo-
therapy in three additional settings (comparisons 2, 5, and
6), bringing the total to seven: (1) surgery versus sur-
gery plus chemotherapy (adjuvant); (2) surgery versus che-
motherapy plus surgery (neoadjuvant); (3) surgery plus
radiotherapy versus surgery plus radiotherapy plus chemo-
therapy; (4) radiotherapy versus sequential radiotherapy
plus sequential chemotherapy; (5) radiotherapy versus ra-
diotherapyplusconcomitantchemotherapy;(6)radiother-
apyplussequentialchemotherapyversusradiotherapyplus
concomitant chemotherapy; (7) supportive care versus
supportive care plus chemotherapy.
For the update of the 1995 meta-analyses, we have
identified a total of 22 new RCTs with more than 8,000
patients in the equivalent setting to that which is de-
scribed by Hotta et al, bringing the total number of trials
to 38. If we can include these patients, it would bring the
total number of patients in this comparison alone to
more than 10,500 patients.
As Piedbois and Buyse point out, IPD meta-analyses
are considered the gold standard but need time and fund-
ing.Thismeta-analysisbyHottaetal,isavaluableresource
in the absence of other evidence, but the results should be
considered with caution until they can be compared with
the updated IPD meta-analysis.
Sarah Burdett and Lesley Stewart
Meta-Analysis Group, MRC Clinical Trials Unit, London, UK
Anne Auperin and Jean-Pierre Pignon
Service de Biostatistique et d’Epidemiologie, Institut Gustave-Roussy,
Villejuif, France
■ ■ ■
Authors’ Disclosures of Potential
Conflicts of Interest
Thefollowingauthorsortheirimmediatefamilymem-
bershaveindicatedafinancialinterest.Noconflictexistsfor
drugs or devices used in a study if they are not being evalu-
ated as part of the investigation. Honoraria: Jean-Pierre
Pignon, Lilly. Research Funding: Jean-Pierre Pignon,
Aventis. For a detailed description of these categories, or
for more information about ASCO’s conflict of interest
policy, please refer to the Author Disclosure Declaration
and the Disclosures of Potential Conflicts of Interest
section of Information for Contributors found in the
front of every issue.
REFERENCES
1. Hotta K, Matsuo K, Ueoka H, et al: Role of adjuvant chemotherapy in
patients with resected non-small cell lung cancer: Reappraisal with a
meta-analysis of randomised controlled trials. J Clin Oncol 22:3860-3867,
2004
2. Non-Small Cell Lung Cancer Collaborative Group: Chemotherapy in
non-small cell lung cancer: A meta-analysis using updated data on individual
patients from 52 randomised clinical trials. BMJ 311:899-909, 1995
3. Piedbois P, Buyse M: Meta-analysis based on abstracted data: A step
in the right direction, but only a first step. J Clin Oncol 22:3839-3841, 2004
DOI: 10.1200/JCO.2005.05.303
IN REPLY: We appreciate the constructive comments
fromBurdettetalconcerningtheimportanceofmeta-analysis
usingindividualpatientdata(IPD)inpatientswithallstagesof
non–small-celllungcancer(NSCLC).Themeta-analysiscon-
ducted by the NSCLC Collaborative Group (NSCLC-CG) in
1995hasbeenaveryimportantandhelpfulreferenceforclini-
ciansinvolvedinNSCLCtreatment.1Itsresultshavealsobeen
an important reference for new clinical trials for NSCLC.
Despite our great respect for the NSCLC-CG study, we were
still interested in whether meta-analysis of trials not included
in the NSCLC-CG study would be in accordance with its
resultsbecausemostofthetrialsintheNSCLC-CGanalysis
involved outdated regimens no longer used in NSCLC
treatment. Therefore, we narrowed our objective to clarify-
ing the role of cytotoxic agents, including platinum or uracil-
tegafur,asadjuvantchemotherapy,andlimitedeligibletrialsto
thoseweanalyzed.2
We have no objection to their statement that IPD-
based meta-analysis is more ideal than abstracted data–
based meta-analysis, in terms of to obtaining answers to
more specific clinical questions. We conducted our ab-
stracteddata–basedanalysistoaddressouraforementioned
clinical question because we were not in a position to con-
ductanIPD-basedanalysis.Weareverypleasedtohearthat
an IPD-based meta-analysis is underway, and we eagerly
await the results. We especially look forward to seeing
whether the results are consistent with ours.
Katsuyuki Hotta
Department of Medicine II, Okayama University Medical School,
Okayama, Japan
Keitaro Matsuo
Division of Epidemiology and Prevention, Aichi Cancer Center Research
Institute, Achai, Japan
Hiroshi Ueoka
Department of Medicine II, Okayama University Medical School,
Okayama, Japan
■ ■ ■
Correspondence
www.jco.org
925
Downloaded from jco.ascopubs.org on October 14, 2011. For personal use only. No other uses without permission.
Copyright © 2005 American Society of Clinical Oncology. All rights reserved.