Can biomarkers play a role in the decision about treatment of the clinically negative neck in...

BASIC SCIENCE REVIEW

Jeffrey N. Myers, MD, PhD, Section Editor

CAN BIOMARKERS PLAY A ROLE IN THE DECISION ABOUTTREATMENT OF THE CLINICALLY NEGATIVE NECK INPATIENTS WITH HEAD AND NECK CANCER?

Robert P. Takes, MD, PhD,1 Alessandra Rinaldo, MD,2 Juan Pablo Rodrigo, MD,3,4

Kenneth O. Devaney, MD, JD, FCAP,5 Johannes J. Fagan, MBChB, FCS (SA) MMed,6

Alfio Ferlito, MD, DLO, DPath, FRCS, FRCSEd, FRCSGlasg, FRCSI, FHKCORL,FDSRCS, FRCPath, FASCP, MCAP2

1 Department of Otolaryngology–Head and Neck Surgery, Radboud University Nijmegen Medical Center,P. O. Box 9101, 6500 HB Nijmegen, The Netherlands. E-mail: [email protected] Department of Surgical Sciences, ENT Clinic, University of Udine, Udine, Italy3 Department of Otolaryngology, Hospital Universitario Central de Asturias, Oviedo, Spain4 Instituto Universitario de Oncologıa del Principado de Asturias, Oviedo, Spain5 Department of Pathology, Foote Hospital, Jackson, Michigan6 Division of Otolaryngology, Faculty of Health Sciences, University of Cape Town, Cape Town, South Africa

Accepted 31 August 2007Published online 22 January 2008 in Wiley InterScience (www.interscience.wiley.com). DOI: 10.1002/hed.20759

Abstract: For most patients with head and neck squamous cell

cancer, a treatment decision concerning the neck is required.

Since detection of small metastases in lymph nodes is difficult,

many of the patients with no detectable metastases receive

elective treatment of the neck. Additional information on the met-

astatic potential of the primary tumor before treatment may be

useful to reduce the number of these elective procedures. Bio-

markers may supply such information. Molecules involved in

several pathways have been studied, but the complexity of the

metastatic process makes it unlikely that a single marker for me-

tastasis can de identified. Techniques allowing the study of

many factors simultaneously seem to be the most promising

ones. In recent years, microarray expression profiling and com-

parative genomic hybridization studies have yielded interesting

results. If these results can be confirmed in larger studies, they

may play a role in future clinical decision making on treatment of

the clinically N0 neck. VVC 2008 Wiley Periodicals, Inc. Head

Neck 30: 525–538, 2008

Keywords: biomarkers; metastasis; neck dissection; head and

neck; cancer

One of the basic issues in the treatment ofpatients with head and neck squamous cell cancer(HNSCC) is the treatment of the clinically N0neck. If nodal metastases are present, there is gen-eral agreement that the neck should be treated; onthe other hand, in the absence of nodal metastasis,the decision to treat the neck is less clear-cut. Theproblem in clinical practice is that the accuracy ofdiagnostic (imaging) techniques to assess the neckis still limited.1,2 Therefore, uncertainty about thetrue status of the neck remains even after the useof the most advanced imaging techniques.

Correspondence to: R. P. Takes

VVC 2008 Wiley Periodicals, Inc.

Role of Biomarkers in the Decision about Treatment of the Neck HEAD & NECK—DOI 10.1002/hed April 2008 525

As a result of the limitations to accurate stag-ing of the neck, many head and neck surgeonselectively treat the neck of patients whose diseasehas been classified N0 when, based on T classifica-tion and site of the primary tumor, the likelihoodof occult metastasis is estimated to be more than15% to 20%. In most patients with an N0 neckwho are treated surgically, a selective neck dissec-tion will be performed.3,4

The alternative strategy for the N0 neck is to‘‘wait and watch.’’5 No study has convincinglydemonstrated higher cure rates for patientsundergoing elective neck dissection when com-pared to ‘‘watchful waiting.’’ However, in the lat-ter group, much will depend on delay of treatmentonce the (previously occult) nodal metastasesbecome apparent, and the intensity of follow-up islikely to influence outcome. Many (chiefly retro-spective) studies have examined this subject withcontroversial results, and so the debate on thetreatment of the N0 neck remains unresolved.

As a result of these conflicting strategies, somepatients with head and neck cancer may receivesuboptimal treatment. As many as 80% of patientswith N0 disease will receive unnecessary treat-ment of the neck (along with its concomitant mor-bidity). Shoulder dysfunction in particular can be adisabling consequence of neck dissections, evenwhen the spinal accessory nerve is preserved, withserious impact on the quality of life.6–8 Moreover,selective neck dissection is often considered a lessappropriate treatment than a comprehensive(modified) radical neck dissection in patients withnodal metastasis, which had been mistakenly clas-sified as N0 before treatment. On the other hand, awatchful waiting strategy harbors the risk of pro-gression of occult metastasis to potentially incura-ble disease. Moreover, extracapsular spread, aknown prognostic factor,9–11 is more common oncemetastasis becomes clinically apparent comparedwith clinically occult metastasis.12

If diagnostic techniques to assess the clinicallynegative neck before treatment could be improved,then metastases that were previously occult wouldbe detected more readily. Such patients would thenbe treated therapeutically instead of electively, andthis would reduce the risk of occult regional metas-tases in the remaining patients with N0 disease. Ifthe incidence of occult regional metastasis couldsignificantly be reduced to<20%, then the numberof elective neck treatments could be decreased.

Most commonly used imaging modalities forassessment of nodal metastases are limited by thesimple fact that metastases need to reach a cer-

tain size before they are detected. Up to one thirdof nodal metastases in laryngeal and hypophar-yngeal cancer are micrometastases (<3 mm),13

which is at best about the detection threshold ofadvanced imaging techniques. This limitation hasled to the search for additional more sensitive pa-rameters or markers for nodal metastasis.

The choice of treatment and the prediction oftumor behavior is currently based mainly on theTNM classification. This in turn is based on ana-tomical considerations. However, tumor behaviormay differ widely within these T categories. Bio-markers reflecting molecular alterations of theprimary tumor and surrounding stroma maysomeday yield information about the metastaticpotential of the tumor. Suitable biomarkers mighteventually enable pathologists and clinicians toestimate the chance of nodal metastases in theindividual patient, irrespective of the actual sizeof the tumor. Such advances would permit moreappropriate therapeutic selection and limit elec-tive treatment of the neck to patients who arelikely to develop nodal metastases based on themolecular profile of the primary tumor.

The processes of carcinogenesis and metasta-sis in particular are bewilderingly complex. Tometastasize, cells have to go through a series ofalterations and modifications.14,15 This meansthat it is very likely that more than 1 marker willbe required to assess an individual patient’s riskof nodal metastasis. The prevailing concept isthat, in the initial phases of tumor progression,tumor cells undergo genetic changes, providingsome kind of proliferative advantage. Theseadvantages may be the ability to resist growth-in-hibiting signals, the avoidance of programmedcell death (apoptosis), and the induction of blood-vessel growth (angiogenesis). To be capable tometastasize, tumor cells must acquire other addi-tional capacities like degradation of the basalmembrane, loss of cell adhesion and migration,lymphangiogenesis, and the ability to survive inthe environment of the metastatic site. Whetherthese metastatic capacities are acquired later orinitially in tumor progression remains unclear,although it seems likely that, in contrast to whathas been commonly presumed, these changes mayoccur early.15 Moreover, many of these capacitiesmay vary in time. For example, cell adhesion mustdecrease to allow cells to migrate and metastasize,but cell adhesion is again needed to settle at themetastatic site.

Many biomarkers have been studied for corre-lations with the presence of nodal metastasis in

526 Role of Biomarkers in the Decision about Treatment of the Neck HEAD & NECK—DOI 10.1002/hed April 2008

HNSCC with widely varying results. Some of theparticularly relevant markers are discussed inthe following sections and are summarized inTable 1. The expression of most of them has beenstudied by immunohistochemistry.

HISTOPATHOLOGICAL FEATURES

Histopathological features like differentiationand tumor-associated inflammatory reaction wereamong the first markers to be studied in relationto the likelihood of development of lymph nodemetastasis in HNSCC.

The degree of inflammatory reaction sur-rounding a tumor has been proposed as a signifi-cant indicator of likelihood of progression of tumorgrowth and metastasis. A correlation between thepresence of an inflammatory reaction and the ab-sence of lymph node metastasis has been reportedin the literature.16

Some authors have described a relationshipbetween the development of lymph node metasta-ses and grade of differentiation.16,17 Poorly differ-entiated tumors should, intuitively, be more likely

to behave aggressively (and carry with them agreater metastatic potential) compared with well-differentiated tumors. However in many studies,different types of grading systems were used,making the evaluation of the results of these stud-ies difficult, and in many studies, a correlationwith the presence of nodal metastases could not beestablished.

The growth pattern of tumors has also beenstudied in relation to tumor behavior. A moreinvasive growth pattern is often presumed to bean indicator of more aggressive tumor behaviorassociated with a greater likelihood of recurrenceand metastasis. In early-stage cancer of the oraltongue, growth pattern and depth of invasionhave indeed been reported to correlate with boththe presence18 and delayed development19–22 ofnodal metastasis.

GENES INVOLVED IN CELL CYCLE REGULATION,

CELL PROLIFERATION, AND APOPTOSIS

Although genes and proteins involved in these ba-sic processes of tumor growth may not seem, on ini-tial examination, to be directly involved in the pro-cess of metastasis, they are nevertheless importantmarkers. For example, many tumor cells undergoapoptosis after intravasation.23 Therefore, if apo-ptosis is inhibited, the metastatic potential mightincrease. Moreover, the tumor cells that arrive at asecondary site do not necessarily proliferate imme-diately. They may remain dormant until conditionsfor growth become favorable.24

Several studies reported that amplification ofthe 11q13 region correlates with a variety of clini-copathological factors, including the presence oflymph node metastasis.16,25–28 Although variousgenes are consistently present on the amplified11q13 region, only CCND1 and EMS1 are likely tobe the key genes within this region, because inaddition to being frequently coamplified, theywere found to be overexpressed in all tumors with11q13 amplification. Cyclin D1 expression, stud-ied with immunohistochemistry, has been foundin about 45% to 50% of HNSCC29 and appears tohave prognostic significance.30 A relation bet-ween overexpression of the cyclin D1 protein andlymph node metastasis has been described inHNSCC.31,32 Other authors, however, did not findsuch a correlation of cyclin D1 expression with Nclassification.29,33

Cortactin, the less frequently studied productof the other 11q13 gene EMS1, is a component ofinvadopodia. These structures are associated with

Table 1. Steps in the process of metastasis and

their relevant molecules.

Cell-cycle regulation, cell proliferation, and apoptosis

Cyclins

EGFR

P53

P21

Survivin

Cell adhesion

CD44 isoforms

Syndecan-1

E-selectin/sialyl Lewis-X and -A

E-cadherin

FAK

Proteolysis (degradation of the basal membrane and

extracellular matrix components)

Matrix metalloproteinases (MMPs)/tissue inhibitors of

metalloproteinases (TIMPs)

Urokinase type plasminogen activators (uPA)/plasminogen

activator inhibitors (PAI)

Cortactin

Microenviromental factors: angiogenesis, lymphangiogenesis,

and hypoxia

Hypoxia inducible factor-1 (HIF-1)

Carbonic anhydrases (CA)

Vascular endothelial growth factor (VEGF)

LYVE-1

D2-40

Chemokines (direction of tumor cells)

CXCL12 (stromal cell-derived factor-1, SDF-1) and its

receptor CXCR4

CCR6 and CCR7

Abbreviations: EGFR, epidermal growth factor receptor; FAK, focal ad-hesion kinase.


extracellular matrix degradation, a topic that willbe discussed later in this article. It may explainthe association between cortactin overexpressionand invasive and aggressive tumor behavior.34

Therefore, although amplification of the 11q13genes seems to be correlated with the presence ofnodal metastasis, this correlation has not conclu-sively been established for the expression of thesegenes.

The epidermal growth factor receptor (EGFR)is a transmembrane glycoprotein that regulatescell growth and invasive behavior. Expression ofEGFR is found in a high number of patients withHNSCC.35 Overexpression of EGFR (measured byimmunohistochemistry) has been associated withenhanced invasive and metastatic capacity, byway of stimulation of matrix metalloproteinases(MMPs).36 A correlation between overexpression(measured by immunohistochemistry or reversetranscriptase-polymerase chain reaction [RT-PCR]35) of EGFR and nodal metastasis35–37 orhigh tumor stage at initial diagnosis38 has beendescribed by some authors, although others havefailed to find such a relation in immunohistochem-ical studies.39,40 However, as the expression rateof EGFR in HNSCC is high, the receptor is atarget for many immunotargeting treatmentstudies,41 and molecular blockade of EGFR mayultimately prove its ability to influence the inva-sive and metastatic capacity in HNSCC.42

One of the most frequently studied markers inHNSCC is p53. Point mutations of p53 are one ofthe most frequent genetic alterations in HNSCC.Such mutations lead to nuclear accumulation ofthe protein. Clinicopathological studies of altera-tions in p53 in HNSCC have shown varyingresults. A correlation with nodal metastasis wasfound in some predominantly immunohistochemi-cal studies43 but was not established in most otherstudies.33,44–48 It appears at this juncture thatp53 is not a significant factor in tumor behavioronce invasion has occurred.48,49

The p21 gene located at 6p21 is transcribed byp53 and produces a cyclin-dependent kinase in-hibitor that induces cell cycle arrest. A correla-tion between decreased expression and nodal me-tastasis was described in some immunohisto-chemical studies,43,50 while others failed to find acorrelation.47

Survivin is a member of the inhibitor of apopto-sis protein family that has been implicated in bothcontrol of cell division and inhibition of apoptosis.Specifically, its antiapoptotic function seems to berelated to the ability to directly or indirectly in-

hibit caspases. Survivin is selectively expressed inthe most common human neoplasms and appearsto be involved in tumor cell resistance to someanticancer agents and ionizing radiation.51 Inaddition, survivin overexpression has been associ-ated with nodal metastasis in laryngeal52 and oro-pharyngeal53 squamous cell carcinoma in immu-nohistochemical studies.

CELL ADHESION

Changes in cellular adhesion are one of the keyprocesses in metastasis; weakening of cell–celland cell–extracellular matrix adhesion is obvi-ously imperative for tumor cells to metastasize.Several molecules involved in cell adhesion havebeen studied, including cadherins, integrins,selectins, and CD44 isoforms.

Cell–cell adhesion in epithelial tumors islargely modulated by cadherins, and E-cadherinin particular. Cadherins are transmembrane gly-coproteins responsible for Ca-dependent cell ad-hesion. Reduced expression can be caused by lossof E-cadherin expression, redistribution to othersites within the cell, and shedding and competi-tion from other proteins.54 Such reduced expres-sion may lead to invasion and metastasis.55 It hasbeen demonstrated that downregulation of the E-cadherin gene is associated with poorly differenti-ated tumors, and invasion and metastasis in a va-riety of different types of cancer.54,56 In several,predominantly immunohistochemical, studies ofHNSCC, a correlation57–59 was found betweenloss of expression of E-cadherin in the primary tu-mor and development of nodal metastases. How-ever, other studies of HNSCC patients failed tofind such a statistically significant relation.60,61

As transmembrane glycoproteins, CD44 isoformsmediate contact between the extracellular matrixand the cytoskeleton. Their expression has beenstudied by immunohistochemistry by several groups.Decreased expression of CD44v6 in HNSCC62 and ofCD44h in laryngeal cancer63 and cancer of the oraltongue64 has been associated by some authors with ahigher metastatic tendency. Other authors, however,found no such a relation for CD44 in oral cancer,65

CD44v6 in supraglottic laryngeal cancer66 and oro-pharyngeal cancer,67 and CD44h in cancer of the lar-ynx and hypopharynx.61,66

Syndecan-1 is a member of the syndecan fam-ily of cell surface heparan sulfate proteoglycans.High expression of this molecule is found on theentire surface of cells of stratified squamous epi-thelia, suggesting a role in cell adhesion and inter-


action with the extracellular matrix. The expres-sion of syndecan-1 is reduced in cancers, includingHNSCC. Differences in expression of syndecan-1have been found to be related to prognosis, clinicalstage of the disease, and the presence or absenceof lymph node metastasis in laryngeal cancer68,69

and other HNSCC,70 although this correlation(with metastasis in particular) was not consis-tently found in all similar immunohistochemicalstudies.71

To metastasize through lymph or blood ves-sels, cancer cells need to bind to endothelial cells.These endothelial cells express several adhesionmolecules. E-selectin, in particular, is such anadhesion molecule that seems to contribute tobinding of cancer cells. It is an inducible cell ad-hesion molecule found on vascular endotheliumat sites of inflammation. Sialyl Lewis-X and -A(sLx and sLa) are its ligands. It is not completelyclear whether expression of sLx and sLa bytumor cells or expression of E-selectin at the siteof the metastasis is an important factor in themetastatic process.72,73

Expression of sLx and sLa on tumor cells hasbeen correlated with advanced tumor stage, sur-vival, and metastatic potential. In patients withskin cancer, expression of sLx and sLa was notfound in normal skin and basal cell cancer, but itwas found in all investigated squamous cell can-cers, perhaps explaining the relatively low meta-static potential of basal cell cancers. Expression ofsLa, but not sLx, was found to correlate withnodal metastasis in oral cancer.73 Other studiesfound a trend toward higher expression of sLx atthe site of metastasis, but no correlation with dis-ease status.72

In addition to reduced cell–cell adhesiveness,loss of adhesion of the epithelial cells to the extrac-ellular matrix is also one of the fundamental path-ways that promote tumor cell migration, invasion,and metastasis. A key factor involved in the con-trol of cellular–extracellular matrix interactionsis the focal adhesion kinase (FAK), an intracellu-lar tyrosine kinase protein that is localized tocellular focal contact sites.74 FAK protein over-expression has been demonstrated in differenttumor types, and in some of them it has been cor-related with the invasive potential of a tumor andpoor patient prognosis.75 FAK has been reportedto be overexpressed in HNSCC, and strong FAKexpression has been associated with nodal metas-tasis.76 Moreover, the combination of E-cadherinand FAK expression resulted in a superior accu-racy to assess the likelihood of nodal metastasis.77

Since invasion and metastasis are very compli-cated multistep processes, it is not surprising thatin this study, by using 2 markers, the ability topredict the presence of nodal metastasis wasincreased when compared with studies that ana-lyzed only E-cadherin or FAK expression.

PROTEOLYSIS

Tumors need enzymes to degrade the basementmembrane and extracellular matrix and to invadesurrounding tissues. Loss of basement membranecomponents, in particular, has been shown to cor-relate with invasive and metastatic potential.78,79

MMPs are enzymes with proteolytic activityagainst extracellular matrix components like col-lagen, elastin, fibronectin, and gelatin. Both tu-mor cells and stromal cells can secrete MMPs.They belong to a family of 21 members. MMP-2and MMP-9, in particular, are capable of degrad-ing collagen type IV, a major component of thebasement membrane. Increased MMP levels havebeen detected in HNSCC and play an importantrole in the invasion and metastasis of squamouscell carcinoma.35,80–83 Tissue inhibitors of metal-loproteinases (TIMPs) inhibit the action of theseMMPs. Thus far, 4 different TIMPs have beenidentified. The balance between MMPs andTIMPs seems to be a key mechanism in the pro-cess of invasion and metastasis. For example,expression of MMP-2 and TIMP-2 have beenfound to correlate with nodal status in SCC of theoral tongue,82 and expression of MMP-1, -2, -3,and -9 and TIMP-1 was found to correlate withnodal metastasis in another immunohistochemi-cal study on oral cancer.81 However, the preciseinteractions of MMPs and TIMPs in tumor inva-sion and metastasis need clarification.

It has been proposed that binding of urokinasetype plasminogen activators (uPA) to its receptoris a mediator of cancer invasion and metastasisthrough degradation of the extracellular matrixby activating MMPs. Inhibitors like plasminogenactivator inhibitors (PAIs) regulate this process.Indications that the uPA system plays a signifi-cant role in the invasive and metastatic potentialof HNSCC were uncovered in several investiga-tions.84,85 Expression of uPA correlated with modeof invasion and secondary development of regionalmetastasis after successful initial treatment forthe primary tumor in a study on oral cancer.85

However, expression of PAI-1 and -2 showed nocorrelation with clinicopathological parameters inthe same study.


TUMOR MICROENVIRONMENT

Angiogenesis and Hypoxia. Several studies havedemonstrated that microvascular density86–88

and hypoxia89 play a role in the process of tumorprogression and metastasis.

Small tumor deposits (up to 1–2 mm in diame-ter) can receive nutrition by diffusion. For furthergrowth, angiogenesis is necessary. Induction ofangiogenesis is mediated by molecules released byboth tumor and host cells.90 Neovascularization isimportant not only for the growth of the primarytumor but also for the access to blood vessels bytumor cells that are destined to metastasize. Met-astatic tumor cells that arrive at a secondary sitedo not necessarily proliferate immediately. Theymay remain dormant until conditions for growthhave become favorable by, for example, angiogene-sis.24 Tumor vascularization is often assessed bymicrovessel density measurements, using anti-bodies against CD-31 or factor VIII related endo-thelial antigen.

The relationship between angiogenesis and tu-mor growth and hematogenic metastasis may seemmore immediately apparent than the relationshipof angiogenesis with lymph node metastasis. How-ever, a correlation between angiogenic activity andthe nodal metastasis has been suggested by severalstudies in different types of cancers.91 Immunohis-tochemical studies in HNSCC have yielded conflict-ing results. Some did reveal such a direct correla-tion86,88,92 while others did not (in smaller, T1 andT2, oral cancers in particular).44,93,94

Tumor hypoxia is an adverse prognostic factor,as it influences tumor progression and response totreatment. Many genes involved in controlling tu-mor biology, including metastatic behavior, areregulated by oxygen. The most important knownoxygen-regulated genes include hypoxia induciblefactor-1 (HIF-1), carbonic anhydrases (CA) (ofwhich CA-9 is the most relevant), genes involvedin apoptosis (p53), and vascular endothelialgrowth factor (VEGF).92 Hypoxia induces in-creased genomic instability in tumors and mayinduce expression of gene products involved in themetastatic process.95

Of the hypoxic markers, HIF-1 is considered tobe a major regulator of physiological responses tohypoxia. It activates the transcription of genesthat are involved in important aspects of tumorbiology, including energy metabolism, angiogene-sis (VEGF), and invasion (MMP).96 Many studieson tumor hypoxia and prognosis have been per-formed in patients treated with radiotherapy,97,98

and a correlation with poor outcome has beenfound.99 In studies on surgically treated patientsusing HIF-1 antibodies, a trend toward a higherincidence of metastasis was found but was not sig-nificant.100 In a study using lactate concentrationas a parameter for hypoxia, it was found to beassociated with development of nodal and distantmetastasis.101

In other tumor types, hypoxic tumors appearto have a higher propensity for nodal metasta-sis.102 Specifically, in HNSCC, it has been sug-gested that hypoxia, measured by CA-9 expres-sion, is associated with higher rates of (distant)metastasis.101 It is interesting in this regard thatoverexpression of CA-9 was found to reduce theE-cadherin–mediated cell–cell adhesion.103 It maybe that either hypoxia or lowered pH is responsi-ble for these effects.101

VEGF is probably the most important angio-genic growth factor in human tumors. It inducesendothelial cell proliferation and is therefore im-portant in wound healing, tumor growth, and me-tastasis. Its expression is increased in hypoxia. Itsclinical significance is not yet clear. In HNSCC, itsexpression has not to date been found to be ofprognostic value.104 However, in recent studies,VEGF has been shown to be a valuable preopera-tive biomarker to predict cervical metastasis inpatients with early oral cancer.105 Yet, theseresults were not confirmed in hypopharyngealcancer.106 Therefore, this biomarker is potentiallyvaluable to guide the treatment of the N0 neck inearly oral cancer. In addition, VEGF expressionalso correlated with survival in oral cancer. Micro-vessel density was also assessed by the expressionof CD105 (endoglin) in these studies, and it wasconsistently associated with lymph node metasta-sis and poor prognosis.105,106

Lymphangiogenesis. The process of tumor cell dis-semination via the blood with the aid of angiogen-esis seems to be better understood than the pro-cess of tumor cell spread through the lymphatics.As is the case with the process of angiogenesis, ithas become apparent that tumors possess a simi-lar ability to induce lymphangiogenesis.

Proliferation of lymphatic vessels is regulatedby members of the VEGF family, specifically byVEGF-C, together with its major receptor VEGFR-3. However, as mentioned previously, VEGF is alsoexpressed on blood vessels. LYVE-1 is a recentlyidentified lymphatic marker, expressed on lym-phatic endothelium.107 High lymph vessel densitywas found to correlate with the presence of


nodal metastasis in HNSCC.107,108 Expression ofother markers for lymphangiogenesis, such as D2-40, was also associated with an increased risk forlymph node metastasis in patients with an accom-panying high degree of lymphangiogenesis.109

Other new and promising endothelial markersfor lymphatic spread may be MR and CLEVER-1.110

CHEMOKINES AND CHEMOKINE RECEPTORS

Chemokines are recognized as important factorsinvolved in tumorigenesis and metastasis in sev-eral types of tumor including HNSCC. The chemo-kine CXCL12 (stromal cell-derived factor-1) andits receptor CXCR4 play important roles in pro-cesses such as inflammation (acting as a chemoat-tractant for leukocytes). Similarly, tumor cells canrespond to chemoattractants produced by certaintissues.111 CXCR4 has recently been character-ized in several types of cancer and is thought toplay a role in directing the metastasizing cells toCXCL12-rich tissues. Expression of CXCR4,measured by RT-PCR and immunohistochemistry,in primary tumors with lymph node metastasiswas significantly higher than in those tumorswithout metastasis.112,113

In HNSCC cells lines arising from the oral cav-ity, CXCR4 was shown to be expressed, whileCXCL12 was found to induce migration of thesecells and to stimulate pathways important in cellmotility. Furthermore, CXCL12 seems to stimulatethe adhesion of HNSCC cells to fibronectin and col-lagen and to promote the activation of MMP-9secreted from HNSCC cells. It so contributes to thedegradation of the basal cell membrane.114

Downregulation of chemokine receptor CCR6and upregulation of CCR7 was consistently foundin metastatic HNSCC and seem to have an impor-tant role in the direction of metastatic cells towardlymphoid tissues.115,116 It may be a promising tar-get for therapy. Correlation of expression of CCR7with nodal metastasis has not yet been studied, toour knowledge.

OTHER MARKERS

Several other putative markers for metastasishave been studied.

eIF-4E may play a role in the survival of meta-static cells by supporting growth factor autonomy.It enhances the translation of proteins involved inthe promotion of several aspects of tumor progres-sion like cell survival and proliferation (c-myc,cyclin D1), angiogenesis (VEGF), and invasion

and metastasis (MMP-9), and may therefore con-tribute to metastasis. Indeed, increased expres-sion can be found in metastasizing tumor cells.117

Squamous cell carcinoma related oncogene(SCCRO) is a gene located on the chromosome3q26.3 that is frequently amplified in HNSCC.118

SCCRO-transfected cells were found to be capableof creating invasive tumors and developing re-gional lymph node metastasis in nude mice.Amplification and overexpression of the SCCROgene correlates with nodal metastasis andoutcome.

Amplification of cyclin L1, located on chromo-some 3q and involved in G0 to G1 transition, cor-related with the presence of nodal metastasis in alarge series of HNSCC.119

DNA PLOIDY

DNA ploidy reflects the amount of genetic mate-rial in the cell nucleus. Most cells are in the rest-ing (G0), diploid phase in normal tissue. Tumorcells usually show genetic instability, and so theDNA content of tumor cells can differ significantlyfrom the normal diploid state. In reports on theDNA ploidy status in HNSCC, correlations havebeen constructed with clinical parameters such asstage and metastasis. In studies of HNSCC of dif-ferent sites examined together, correlation withstage or nodal metastasis was found in someseries,45,120,121 but in others such a correlationcould not be established.61,122 A correlation123–125

or some relation126 with lymph node metastasis incancer of the oral cavity has been reported.

HIGH THROUGHPUT MARKER STUDIES FOR

METASTASIS IN HNSCC

If markers are to be used for clinical purposes, thetechniques to detect them should ideally be read-ily available and relatively easy for the clinicallaboratory to perform. Most of the aforementionedmarkers can be studied on a protein level byimmunohistochemistry. These techniques are rel-atively easy to execute and are relatively inexpen-sive. However, they are less practical when testinglarge panels of markers. Since a combination ofmultiple markers will probably be a better predic-tor of behavior than single markers alone, techni-ques allowing study of numerous markers simul-taneously would be very attractive.

More recently, high throughput techniqueshave been developed. These techniques includegene-expression profiling and comparative genomic


hybridization (CGH) with microarray technology,which permit the study of (ten-) thousands of genessimultaneously. The resulting genetic signaturesmay be useful as predictors of the likelihood of exhi-bition of metastatic behavior, and may offer insightinto the underlying processes and moleculesinvolved in these processes.

A number of microarray gene expression stud-ies of metastasis in HNSCC have been performedin recent years. Many of these studies comprisedonly a limited number of cases and should there-fore be considered no more than pilot studies.127–130

However, Chung et al131 studied 60 HNSCCusing cDNA microarrays and used a supervisedmethod of analysis to predict the N classificationof primary tumors. They were able to accuratelypredict the presence of nodal metastasis in 80% ofpatients. Roepman et al132 analyzed 82 primarytumors located in the oropharynx or oral cavitywith oligonucleotide arrays and found a set of 102genes that was able to predict the presence of me-tastasis in 74% of patients, with improved resultswith more recently stored samples. Furthermore,in this work, the predictor has a predictive accu-racy for N0 status of 100%. No false-negative pre-dictions were made, which is most important forthe goal of achieving clinical relevance. Genes inthis classifying set included epithelial markergenes, genes encoding extracellular matrix com-ponents, genes involved in cell adhesion, celldeath, cell growth, and extracellular matrixremodeling. The initial set of 102 predictive genesis a subset of a larger group of 825 genes with pre-dictive power. In a further analysis133 of the com-position of the metastatic signature, the authorsfound that many of the predictive genes wereinterchangeable because of a similar expressionpattern across the tumor samples, and raising thenumber of genes included in the signature couldcompensate for exclusion of the strongest predic-tive genes. Multiple accurate predictive signa-tures can be designed using various subsets of pre-dictive genes. Including more genes with lowerpredictive power can compensate for the absenceof genes with strong predictive power. This sug-gests that there are more predictive genes thanrequired to design an accurate predictor.

A few CGH studies studying differences inDNA copy numbers between N0 and N1 primarytumors have been performed. Bockmuhl et al,134

using conventional CGH, studied differencesbetween primary tumors and their correspondingmetastases, as well as between metastasizing andnonmetastasizing primary HNSCC. They found

that deletion of regions 5q34-q35, 8p12-p22,10p12, 10q21-qter, 11p14-p15, 11q14, 11q23-qter,and 14q21-qter, and overrepresentation of 1q21-q22, 3q24-qter, 6q, 7q11.2, 12q12-q14, and18p11.2, were significantly associated with meta-stasizing tumors. Chen et al135 used microarrayCGH to perform a genome-wide screening forchromosomal aberrations in 60 oral squamous cellcarcinomas. When comparing these oral cancerswith and without nodal involvement, they identi-fied 12 targets that appeared to be associated withthe presence of lymph node metastasis: loss of 18qtel, COX2 at 1q31.1 and INSR at 19p13.2, andgain of PRKCZ at region 1p36.33, region 6p tel,region 7p tel, EGFR at 7p12.1-3, CCND1, EMS1and FGF4/FGF3 at 11q13, THRA at 17q11.2, andAIB1 at 20q12.

COMMENTS ON THE USE OF BIOMARKERS FOR

CLINICAL PURPOSES

Although the use of biomarkers may seem verypromising, there are several critical factors yet toconsider. If markers are to be used in decisionmaking with regard to treatment of the neck inpatients with head and neck carcinoma, the onlyavailable material to study before treatmentwould be biopsy material. This biopsy materialmay not be representative of the tumor in its en-tirety as a consequence of the heterogeneity oftumors.136 Moreover, some traditional light micro-scopic parameters like depth of invasion may bemore difficult to assess in biopsy material than inthe resection specimen. Markers reflecting theinteraction between tumor and surroundingstroma may also be more difficult to study on bi-opsy material. Some of these objections may be cir-cumvented by a 2-stage surgical procedure withfirst resection and examination of the primary tu-mor, and subsequently a decision on the treatmentof the neck based on the examination of the fea-tures of the resected specimen. Another factorrelated to the heterogeneity of tumors is the factthat changes in genes and protein expression mayvary from 1 part of a given tumor to another partof the same tumor. In the course of tumor progres-sion, some cells of a clone may acquire additionalor different chromosomal alterations from theirneoplastic neighbors, resulting in a subclone withdifferent properties. The question then is whichpercentage of cells showing expression of a certainmarker should determine a positive scoring resultand which percentage a negative one. As could bepredicted, different cutoff points have been


selected in different studies. More uniformity isrequired in scoring categories, if markers are to beused for clinical purposes.

Instead of trying to characterize the entire tu-mor, it may be even more important to detectsmall subclones in a tumor with genetic changes,which are, presumably, relevant for metastasis. Intheory, only a small portion of the tumor can meta-stasize. However, some authors have found thatthe metastatic potential of tumors is encoded inthe bulk of a primary tumor, thus challengingthe previously noted belief that metastases arisefrom rare cells with a capacity for metastasizing,within a larger primary tumor.137

Another factor to take into consideration is thepossibility that different techniques can be used tostudy the same features of a tumor. Proteinexpression, for example, is studied in differentways by different authors. Techniques likeimmunofluorescence, immunohistochemistry, andWestern blotting have been used for this purpose,making comparison of results of different studiesdifficult. Some authors use fresh-frozen material,whereas others use paraffin-embedded material.Moreover, many different antibodies directedagainst the same protein of interest are oftenavailable, and different staining protocols are fol-lowed. This may also contribute to some variationin results.

No distinction is made between primary tumorsubsites in many studies of HNSCC. It may bethat results derived from tumors arising in 1locale are not immediately applicable to tumorsarising in other locales. There are indeed indica-tions of differences in expression of severalgenetic markers between tumors arising in thedifferent subsites of the head and neck.138–140

These differences are probably reflected in thewell-known variation of biological behavior oftumors arising in different areas of the head andneck. Since some studies are restricted to particu-lar locales of the head and neck and other studiesconsider all these locales together, the results ofthese series are not comparable. This leads to aninconsistency of results and will therefore impedethe acceptance of the use of biomarkers for clinicalpurposes.

There may be problems with the interpretationof the results of marker expression; but the samemay be said with regard to the parameter to whichthis marker expression is related, ie, the N status.That is to say, the means by which this N statuswas determined is not always precisely defined.Histopathological examination of the neck dissec-

tion specimen is the most precise means of deter-mining the status of the neck. However, this goldstandard for the nodal status very much dependson the way the neck dissection specimen is exam-ined.141

In most studies, the nodal status is based on(routine) histopathological examination of theneck dissection specimen. However, small meta-static deposits or micrometastases may still re-main undetected.142 If the lymph nodes are exam-ined more meticulously, small metastases may bedetected, which may not yet be detected by con-ventional light microscopic examination. Specialimmunohistochemical investigations, which en-hance the detection of particularly small depositsof malignant epithelial cells, have found addi-tional micrometastases in 5% to 50% of thepatients who had no evidence of metastases at thetime of routine pathologic examinations, with amean incidence rate of 15.2%.143 The use of molec-ular techniques increases this number yet further,to around 20%, in some studies.144–147 Apart fromthe clinical significance of finding these very smalltumor deposits, it seems clear that the choice ofreference standard for assessment of the cervicalnodal status will influence the outcome of studiesof the identification of markers correlated withnodal metastasis.

Finally, 1 of the most important reasons thatmarkers are not yet used for the evaluation of thenodal status of patients may be that, until theresults of microarray studies were reported, moststudies focused only on single markers. Since theprocess of metastasis is very complex and manyfactors are involved, it is unlikely that a singlemarker is sufficient to predict the metastaticbehavior of a tumor. The complexity of the meta-static process suggests that multiple parametersor markers will be needed.16 For the purpose ofstudying multiple genetic factors, microarrays areideal. However, some points should again be con-sidered concerning microarray studies.

mRNA transcript levels are measured inexpression array studies; while this may be a goodrepresentation of the transcriptional activity of agene, it does not measure the amount of activeprotein in the tissue, as it does not take intoaccount posttranscriptional regulation such asactivation of a protein or the rate at which the pro-tein is broken down. Thus, the outcome will prob-ably have to be validated on other platforms thatmeasure the actual protein, to determine whetherthe transcriptional activity of a gene translates anactual protein activity in the tumor tissue.


Second, expression arrays are performed onsnap-frozen tissue. RNA is fragile and easily bro-ken down, and the time from taking the biopsy tofreezing it is critical as is the time from storage ofthe tissue until analysis.132 Moreover, the isolatedmRNA has to be labeled and amplified before itcan be hybridized to the microarray. During theseprocedures, the RNA is also sensitive to break-down. Furthermore, not all mRNA fragments maybe amplified at the same rate, introducing anotherpotential source of bias in the end result.

Third, the enormous amount of data generatedby a microarray experiment poses a statisticalchallenge that requires new and sophisticatedanalysis methods. The choice of the analysis tech-nique is a crucial element in any microarraystudy, for it can dramatically affect the outcome.

Finally, critics point out that the profiles iden-tified in different studies do not always match upwith one another.148 However, many genes withsimilar functions may be interchangeable, and sopredictive profiles for the same tumor entity maybe composed of different gene sets.133,149

The results are very promising, but now needto be validated on larger numbers of patients. Cur-rently multicenter studies are underway throughthe American College of Surgery Oncology Groupand NWHHT (Dutch Head and Neck CooperativeGroup).

PERSPECTIVES

Although it is unlikely that it will be possible to be100% certain about the true status of the neck,continuing improvements in imaging techniqueswill contribute to improvement of pretreatmentstaging. The number of patients with occultmetastasis will further decrease in this way. Bio-markers may provide additional and complemen-tary information and may be able to identifypatients with a low chance of having occult metas-tasis. This group may be suitable for a wait-and-see policy for the neck. Of the available markersand techniques currently available, microarrayexpression profiling and CGH seem the mostpromising. If the results of the first emergingstudies can be confirmed in sufficiently largeseries of patients, these arrays may be introducedinto clinical practice.

REFERENCES

1. van den Brekel MWM, Castelijns JA, Snow GB. Diag-nostic evaluation of the neck. Otolaryngol Clin NorthAm 1998;31:601–619.

2. Takes RP. Staging of the neck in patients with headand neck squamous cell cancer: imaging techniquesand biomarkers. Oral Oncol 2004;40:656–667.

3. Ferlito A, Rinaldo A, Silver CE, et al. Elective and ther-apeutic selective neck dissection. Oral Oncol 2006;42:14–25.

4. Wei WI, Ferlito A, Rinaldo A, et al. Management of theN0 neck—reference or preference. Oral Oncol 2006;42:115–122.

5. Nieuwenhuis EJ, Castelijns JA, Pijpers R, et al. Wait-and-see policy for the N0 neck in early-stage oral andoropharyngeal squamous cell carcinoma using ultraso-nography-guided cytology: is there a role for identifica-tion of the sentinel node? Head Neck 2002;24:282–289.

6. van Wilgen CP, Dijkstra PU, van der Laan BF, PlukkerJT, Roodenburg JL. Morbidity of the neck after headand neck cancer therapy. Head Neck 2004;26:785–791.

7. Kuntz AL, Weymuller EA Jr. Impact of neck dissectionon quality of life. Laryngoscope 1999;109:1334–1338.

8. van Wilgen CP, Dijkstra PU, van der Laan BF, PlukkerJT, Roodenburg JL. Shoulder and neck morbidity inquality of life after surgery for head and neck cancer.Head Neck 2004;26:839–844.

9. Ferlito A, Rinaldo A, Devaney KO, et al. Prognostic sig-nificance of microscopic and macroscopic extracapsularspread from metastatic tumor in the cervical lymphnodes. Oral Oncol 2002;38:747–751.

10. Greenberg JS, Fowler R, Gomez J, et al. Extent ofextracapsular spread: a critical prognosticator in oraltongue cancer. Cancer 2003;97:1464–1470.

11. Woolgar JA, Rogers SN, Lowe D, Brown JS, VaughanED. Cervical lymph node metastasis in oral cancer: theimportance of even microscopic extracapsular spread.Oral Oncol 2003;39:130–137.

12. Greenberg JS, El Naggar AK, Mo V, Roberts D, MyersJN. Disparity in pathologic and clinical lymph nodestaging in oral tongue carcinoma. Implication for thera-peutic decision making. Cancer 2003;98:508–515.

13. Buckley JG, MacLennan K. Cervical node metastasesin laryngeal and hypopharyngeal cancer: a prospectiveanalysis of prevalence and distribution. Head Neck2000;22:380–385.

14. Liotta LA. Cancer cell invasion and metastasis. Sci Am1992;266:54–59, 62–63.

15. Bernards R, Weinberg RA. A progression puzzle.Nature 2002;418:823.

16. Takes RP, Baatenburg de Jong RJ, Schuuring E, et al.Markers for assessment of nodal metastasis in laryn-geal carcinoma. Arch Otolaryngol Head Neck Surg 1997;123:412–419.

17. Rodrigo JP, Domınguez F, Alvarez C, Manrique C,Herrero A, Suarez C. Expression of E-cadherin in squa-mous cell carcinomas of the supraglottic larynx withcorrelations to clinicopathological features. Eur J Can-cer 2002;38:1059–1064.

18. O-charoenrat P, Pillai G, Patel S, et al. Tumour thick-ness predicts cervical nodal metastases and survival inearly oral tongue cancer. Oral Oncol 2003;39:386–390.

19. Jones AS. Prognosis in mouth cancer: tumour factors.Eur J Cancer B Oral Oncol 1994;30:8–15.

20. Kurokawa H, Yamashita Y, Takeda S, Zhang M,Fukuyama H, Takahashi T. Risk factors for late cervicallymph node metastases in patients with stage I or IIcarcinoma of the tongue. Head Neck 2002;24:731–736.

21. Okamoto M, Nishimine M, Kishi M, et al. Prediction ofdelayed neck metastasis in patients with stage I/IIsquamous cell carcinoma of the tongue. J Oral PatholMed 2002;31:227–233.

22. Lim SC, Zhang S, Ishii G, et al. Predictive markers forlate cervical metastasis in stage I and II invasive squa-mous cell carcinoma of the oral tongue. Clin CancerRes 2004;10:166–172.


23. Wong CW, Lee A, Shientag L, et al. Apoptosis: an earlyevent in metastatic inefficiency. Cancer Res 2001;61:333–338.

24. Luzzi KJ, MacDonald IC, Schmidt EE, et al. Multistepnature of metastatic inefficiency: dormancy of solitarycells after successful extravasation and limited survivalof early micrometastases. Am J Pathol 1998;153:865–873.

25. Schuuring E. The involvement of the chromosome11q13 region in human malignancies: cyclin D1 andEMS1 are two new candidate oncogenes. Gene 1995;59:83–96.

26. Jares P, Fernandez PL, Campo E, et al. PRAD-1/cyclinD1 gene amplification correlates with messenger RNAoverexpression and tumor progression in human laryn-geal carcinomas. Cancer Res 1994;54:4813–4817.

27. Alavi S, Namazie A, Calcaterra TC, Wang MB, SrivatsanES. Clinical application of fluorescence in situ hybridiza-tion for chromosome 11q13 analysis in head and neckcancer. Laryngoscope 1999;109:874–879.

28. Rodrigo JP, Garcia LA, Ramos S, Lazo PS, Suarez C.EMS1 gene amplification correlates with poor prognosisin squamous cell carcinomas of the head and neck. ClinCancer Res 2000;6:3177–3182.

29. Michalides RJAM, van Veelen NMJ, Kristel PMP, et al.Overexpression of cyclin D1 indicates a poor prognosisin squamous cell carcinoma of the head and neck. ArchOtolaryngol Head Neck Surg 1997;123:497–502.

30. Yu Z, Weinberger PM, Haffty BG, et al. Cyclin D1 is avaluable prognostic marker in oropharyngeal squamouscell carcinoma. Clin Cancer Res 2005;11:1160–1166.

31. Masuda M, Hirakawa N, Nakashima T, Kuratomi Y,Komiyama S. Cyclin D1 overexpression in primaryhypopharyngeal carcinomas. Cancer 1996;78:390–395.

32. Capaccio P, Pruneri G, Carboni N, et al. Cyclin D1expression is predictive of occult metastases in headand neck cancer patients with clinically negative cervi-cal lymph nodes. Head Neck 2000;22:234–240.

33. Mielcarek-Kuchta D, Olofsson J, Golusinski W. p53,Ki67 and cyclin D1 as prognosticators of lymph nodemetastases in laryngeal carcinoma. Eur Arch Otorhino-laryngol 2003;260:549–554.

34. Clark ES, Whigham AS, Yarbrough WG, Weaver AM.Cortactin is an essential regulator of matrix metallo-proteinase secretion and extracellular matrix degrada-tion in invadopodia. Cancer Res 2007;67:4227–4235.

35. Kusukawa J, Harada H, Shima I, Sasaguri Y,Kameyama T, Morimatsu M. The significance of epider-mal growth factor receptor and matrix metalloprotei-nase-3 in squamous cell carcinoma of the oral cavity.Eur J Cancer B Oral Oncol 1996;32:217–221.

36. O-charoenrat P, Rhys-Evans PH, Archer DJ, Eccles SA.C-erbB receptors in squamous cell carcinomas of thehead and neck: clinical significance and correlationwith matrix metalloproteinases and vascular endothe-lial growth factors. Oral Oncol 2002;38:73–80.

37. Xia W, Lau YK, Zhang HZ, et al. Combination ofEGFR, HER-2/neu, and HER-3 is a stronger predictorfor the outcome of oral squamous cell carcinoma thanany individual family members. Clin Cancer Res 1999;5:4164–4174.

38. Nguyen DC, Parsa B, Close A, Magnusson B, CroweDL, Sinha UK. Overexpression of cell cycle regulatoryproteins correlates with advanced tumor stage in headand neck squamous cell carcinomas. Int J Oncol 2003;22:1285–1290.

39. Ulanovski D, Stern Y, Roizman P, Shpitzer T, PopovtzerA, Feinmesser R. Expression of EGFR and Cerb-B2 asprognostic factors in cancer of the tongue. Oral Oncol2004;40:532–537.

40. Ang KK, Berkey BA, Tu X, et al. Impact of epidermalgrowth factor receptor expression on survival and pat-

tern of relapse in patients with advanced head andneck carcinoma. Cancer Res 2002;62:7350–7356.

41. O-charoenrat P, Rhys-Evans PH, Modjtahedi H, EcclesSA. The role of c-erbB receptors and ligands in headand neck squamous cell carcinoma. Oral Oncol 2002;38:627–640.

42. Huang SM, Li J, Harari PM. Molecular inhibition ofangiogenesis and metastatic potential in human squa-mous cell carcinomas after epidermal growth factor re-ceptor blockade. Mol Cancer Ther 2002;1:507–514.

43. Tatemoto Y, Osaki T, Yoneda K, Yamamoto T, Ueta E,Kimura T. Expression of p53 and p21 proteins in oralsquamous cell carcinoma: correlation with lymph nodemetastasis and response to chemoradiotherapy. PatholRes Pract 1998;194:821–830.

44. Leedy DA, Trune DR, Kronz JD, Weidner N, Cohen JI.Tumor angiogenesis, the p53 antigen, and cervical me-tastasis in squamous carcinoma of the tongue. Otolar-yngol Head Neck Surg 1994;111:417–422.

45. Pruneri G, Pignataro L, Carboni N, et al. Clinical rele-vance of p53 and bcl-2 protein over-expression in laryn-geal squamous-cell carcinoma. Int J Cancer 1998;79:263–268.

46. Van Heerden WF, Dreyer L, Swart TJ, Van HeerdenMB, Boy SC. The suitability of paraffin-embedded ma-terial to predict metastatic potential of oral squamouscell carcinoma. Anticancer Res 2002;22:4147–4150.

47. Yuen PW, Chow V, Choy J, Lam KY, Ho WK, Wei WI.The clinicopathologic significance of p53 and p21 expres-sion in the surgical management of lingual squamouscell carcinoma. Am J Clin Pathol 2001;116:240–245.

48. Cabanillas R, Rodrigo JP, Astudillo A, Domınguez F,Suarez C, Chiara MD. P53 expression in squamous cellcarcinomas of the supraglottic larynx and its lymphnode metastases: new results for an old question. Can-cer 2007;109:1791–1798.

49. Bosch FX, Ritter D, Enders C, et al. Head and neck tu-mor sites differ in prevalence and spectrum of p53alterations but these have limited prognostic value. IntJ Cancer 2004;111:530–538.

50. Kuo MY, Chang HH, Hahn LJ, Wang JT, Chiang CP.Elevated ras p21 expression in oral premalignantlesions and squamous cell carcinoma in Taiwan. J OralPathol Med 1995;24:255–260.

51. Zaffaroni N, Pennati M, Daidone MG. Survivin as atarget for new anticancer interventions. J Cell Mol Med2005;9:360–372.

52. Marioni G, Bertolin A, Giacomelli L, et al. Expressionof the apoptosis inhibitor protein survivin in primarylaryngeal carcinoma and cervical lymph node metasta-sis. Anticancer Res 2006;26:3813–3817.

53. Marioni G, Bedogni A, Giacomelli L, et al. Survivinexpression is significantly higher in pN1oral and oro-pharyngeal primary squamous cell carcinomas than inpN0 carcinomas. Acta Otolaryngol 2005;125:1218–1223.

54. Jiang WG. E-cadherin and its associated protein cate-nins, cancer invasion and metastasis. Br J Surg 1996;83:437–446.

55. Kudo Y, Kitajima S, Ogawa I, et al. Invasion and me-tastasis of oral cancer cells require methylation of E-cadherin and/or degradation of membranous b-catenin.Clin Cancer Res 2004;10:5455–5463.

56. Takeichi M. Cadherins in cancer: implications for inva-sion and metastasis. Curr Opin Cell Biol 1993;5:806–811.

57. Schipper JH, Unger A, Jahnke K. E-cadherin as a func-tional marker of the differentiation and invasiveness ofsquamous cell carcinoma of the head and neck. ClinOtolaryngol 1994;19:381–384.

58. Eriksen JG, Steiniche T, Sogaard H, Overgaard J.Expression of integrins and E-cadherin in squamouscell carcinomas of the head and neck. APMIS 2004;112:560–568.


59. Franchi A, Gallo O, Boddi V, Santucci M. Prediction ofoccult neck metastases in laryngeal carcinoma: role ofproliferating cell nuclear antigen, MIB-1, and E-cad-herin immunohistochemical determination. Clin CancerRes 1996;2:1801–1808.

60. Andrews NA, Jones AS, Helliwell TR, Kinsella AR.Expression of the E-cadherin-catenin adhesion complexin primary squamous cell carcinomas of the head andneck and their nodal metastases. Br J Cancer 1997;75:1474–1480.

61. Liu M, Lawson G, Delos M, et al. Prognostic value ofcell proliferation markers, tumour suppressor proteinsand cell adhesion molecules in primary squamous cellcarcinoma of the larynx and hypopharynx. Eur ArchOtorhinolaryngol 2003;260:28–34.

62. Kanke M, Fujii M, Kameyama K, et al. Role of CD44variant exon 6 in invasion of head and neck squamouscell carcinoma. Arch Otolaryngol Head Neck Surg2000;126:1217–1223.

63. Spafford MF, Koeppe J, Pan Z, Archer PG, Meyers AD,Franklin WA. Correlation of tumor markers p53, bcl-2,CD34, CD44H, CD44v6, and Ki-67 with survival andmetastasis in laryngeal squamous cell carcinoma. ArchOtolaryngol Head Neck Surg 1996;122:627–632.

64. Masuda M, Kuratomi Y, Shiratsuchi H, Nakashima T,Naonobu K, Komiyama S. Decreased CD44H expressionin early-stage tongue carcinoma associates with latenodal metastases following interstitial brachytherapy.Head Neck 2000;22:662–665.

65. Kosunen A, Pirinen R, Ropponen K, et al. CD44 expres-sion and its relationship with MMP-9, clinicopathologi-cal factors and survival in oral squamous cell cancer.Oral Oncology 2007;43:51–59.

66. Rodrigo JP, Domınguez F, Alvarez C, Gonzalez MV,Herrero A, Suarez C. Clinicopathologic significance ofexpression of CD44s and CD44v6 isoforms in squamouscell carcinoma of the supraglottic larynx. Am J ClinPathol 2002;118:67–72.

67. Kawano T, Nakamura Y, Yanoma S, et al. Expression ofE-cadherin, and CD44s and CD44v6 and its associationwith prognosis in head and neck cancer. Auris NasusLarynx 2004;31:35–41.

68. Klatka J. Syndecan-1 expression in laryngeal cancer.Eur Arch Otorhinolaryngol 2002;259:115–118.

69. Pulkkinen JO, Penttinen M, Jalkanen M, Klemi P,Grenman R. Syndecan-1: a new prognostic marker inlaryngeal cancer. Acta Otolaryngol 1997;117:312–315.

70. Anttonen A, Kajanti M, Heikkila P, Jalkanen M, JoensuuH. Syndecan-1 expression has prognostic significance inhead and neck carcinoma. Br J Cancer 1999;79:558–564.

71. Inki P, Joensuu H, Grenman R, Klemi P, Jalkanen M.Association between syndecan-1 expression and clinicaloutcome in squamous cell carcinoma of the head andneck. Br J Cancer 1994;70:319–323.

72. Farmer RW, Richtsmeier WJ, Scher RL. Identificationof sialyl Lewis-x in squamous cell carcinoma of thehead and neck. Head Neck 1998;20:726–731.

73. Kurahara S, Shinohara M, Ikebe T, et al. Immunohisto-chemical study of sialyl Le(a) and sialyl Le(x) antigenin oral squamous cell carcinoma: the association ofsialyl Le(a) expression with the metastatic potential.Head Neck 1999;21:330–337.

74. Schaller MD, Borgmana CA, Cobb BS, Vines RR, Reyn-olds AB, Parsons JT. pp125FAK, a structurally uniqueprotein tyrosine kinase associated with focal adhesions.Proc Natl Acad Sci U S A 1992;89:5192–5196.

75. McLean GW, Carragher NO, Avizienyte E, Evans J,Brunton V, Frame MC. The role of focal-adhesion ki-nase in cancer—a new therapeutic opportunity. NatRev Cancer 2005;5:505–515.

76. Canel M, Secades P, Rodrigo JP, et al. Overexpressionof focal adhesion kinase in head and neck squamous

cell carcinoma is independent of fak-gene copy number.Clin Cancer Res 2006;12:3272–3279.

77. Rodrigo JP, Dominguez F, Suarez V, Canel M, SecadesP, Chiara MD. Focal adhesion kinase and E-cadherinas a marker for nodal metastasis in laryngeal cancer.Arch Otolaryngol Head Neck Surg 2007;133:145–150.

78. Hirota J, Yoneda K, Osaki T. Basement membrane typeIV collagen in oral squamous cell carcinoma. HeadNeck 1990;12:400–405.

79. Kumagai S, Kojima S, Imai K, et al. Immunohistologicdistribution of basement membrane in oral squamouscell carcinoma. Head Neck 1994;16:51–57.

80. Curran S, Murray GI. Matrix metalloproteinases intumour invasion and metastasis. J Pathol 1999;189:300–308.

81. Kurahara S, Shinohara M, Ikebe T, et al. Expression ofMMPs, MT-MMP, and TIMPs in squamous cell carci-noma of the oral cavity: correlations with tumor inva-sion and metastasis. Head Neck 1999;21:627–638.

82. Yoshizaki T, Maruyama Y, Sato H, Furukawa M.Expression of tissue inhibitor of matrix metalloprotei-nase-2 correlates with activation of matrix metallopro-teinase-2 and predicts poor prognosis in tongue squa-mous cell carcinoma. Int J Cancer 2001;95:44–50.

83. Culhaci N, Metin K, Copcu E, Dikicioglu E. Elevatedexpression of MMP-13 and TIMP-1 in head and necksquamous cell carcinomas may reflect increased tumorinvasiveness. BMC Cancer 2004;4:42.

84. Petruzzelli GJ, Snyderman CH, Johnson JT. In vitrourokinase type plasminogen activator levels and totalplasminogen activator activity in squamous cell carci-nomas of the head and neck. Arch Otolaryngol HeadNeck Surg 1994;120:989–992.

85. Nozaki S, Endo Y, Kawashiri S, et al. Immunohisto-chemical localization of a urokinase-type plasminogenactivator system in squamous cell carcinoma of the oralcavity: association with mode of invasion and lymphnode metastasis. Oral Oncol 1998;34:58–62.

86. Gasparini G, Weidner N, Maluta S, et al. Intratumoralmicrovessel density and p53 protein: correlation withmetastasis in head-and-neck squamous-cell carcinoma.Int J Cancer 1993;55:739–744.

87. Klijanienko J, el-Naggar AK, de Braud F, et al. Tumorvascularization, mitotic index, histopathologic grade,and DNA ploidy in the assessment of 114 head andneck squamous cell carcinomas. Cancer 1995;75:1649–1656.

88. Shpitzer T, Chaimoff M, Gal R, Stern Y, Feinmesser R,Segal K. Tumor angiogenesis as a prognostic factor inearly oral tongue cancer. Arch Otolaryngol Head NeckSurg 1996;122:865–868.

89. Dachs GU, Tozer GM. Hypoxia modulated gene expres-sion: angiogenesis, metastasis and therapeutic exploita-tion. Eur J Cancer 2000;36:1649–1660.

90. Fidler IJ, Ellis LM. The implications of angiogenesisfor the biology and therapy of cancer metastasis. Cell1994;79:185–188.

91. Weidner N, Semple JP, Welch WR, Folkman J. Tumorangiogenesis and metastasis—correlation in invasivebreast carcinoma. N Engl J Med 1991;324:1–8.

92. Williams JK, Carlson GW, Cohen C, DeRose PB,Hunter S, Jurkiewicz MJ. Tumor angiogenesis as aprognostic factor in oral cavity tumors. Am J Surg1994;168:373–380.

93. Zatterstrom UK, Brun E, Willen R, Kjellen E, WennerbergJ. Tumor angiogenesis and prognosis in squamous cellcarcinoma of the head and neck. Head Neck 1995;17:312–318.

94. Gleich LL, Biddinger PW, Pavelic ZP, Gluckman JL.Tumor angiogenesis in T1 oral cavity squamous cellcarcinoma: role in predicting tumor aggressiveness.Head Neck 1996;18:343–346.


95. Rofstad EK. Microenvironment-induced cancer metas-tasis. Int J Radiat Biol 2000;76:589–605.

96. Semenza GL. Targeting HIF-1 for cancer therapy. NatRev Cancer 2003;3:721–732.

97. Dietz A, Vanselow B, Rudat V, et al. Prognostic impactof reoxygenation in advanced cancer of the head andneck during the initial course of chemoradiation orradiotherapy alone. Head Neck 2003;25:50–58.

98. Kaanders JH, Wijffels KI, Marres HA, et al. Pimonida-zole binding and tumor vascularity predict for treat-ment outcome in head and neck cancer. Cancer Res2002;62:7066–7074.

99. Hoogsteen IJ, Marres HA, Wijffels KI, et al. Colocaliza-tion of carbonic anhydrase 9 expression and cell prolif-eration in human head and neck squamous cell carci-noma. Clin Cancer Res 2005;11:97–106.

100. Beasley NJ, Leek R, Alam M, et al. Hypoxia-induciblefactors HIF-1a and HIF-2a in head and neck cancer:relationship to tumor biology and treatment outcome insurgically resected patients. Cancer Res 2002;62:2493–2497.

101. Brizel DM, Schroeder T, Scher RL, et al. Elevated tu-mor lactate concentrations predict for an increased riskof metastases in head-and-neck cancer. Int J RadiatOncol Biol Phys 2001;51:349–353.

102. Hockel M, Vaupel P. Tumor hypoxia: definitions andcurrent clinical, biologic, and molecular aspects. J NatlCancer Inst 2001;93:266–276.

103. Svastova E, Zilka N, Zat’ovicova M, et al. Carbonicanhydrase IX reduces E-cadherin-mediated adhesion ofMDCK cells via interaction with b-catenin. Exp CellRes 2003;290:332–345.

104. Salven P, Heikkila P, Anttonen A, Kajanti M, JoensuuH. Vascular endothelial growth factor in squamous cellhead and neck carcinoma: expression and prognosticsignificance. Mod Pathol 1997;10:1128–1133.

105. Chien CY, Su CY, Hwang CF, Chuang HC, Chen CM,Huang CC. High expressions of CD105 and VEGF inearly oral cancer predict potential cervical metastasis.J Surg Oncol 2006;94:413–417.

106. Chien CY, Su CY, Hwang CF, et al. Clinicopathologicsignificance of CD105 expression in squamous cell car-cinoma of the hypopharynx. Head Neck 2006;28:441–446.

107. Beasley NJ, Prevo R, Banerji S, et al. Intratumorallymphangiogenesis and lymph node metastasis in headand neck cancer. Cancer Res 2002;62:1315–1320.

108. Maula SM, Luukkaa M, Grenman R, Jackson D,Jalkanen S, Ristamaki R. Intratumoral lymphatics areessential for the metastatic spread and prognosis insquamous cell carcinomas of the head and neck region.Cancer Res 2003;63:1920–1926.

109. Franchi A, Gallo O, Massi D, Baroni G, Santucci M.Tumor lymphangiogenesis in head and neck squamouscell carcinoma: a morphometric study with clinical cor-relations. Cancer 2004;101:973–978.

110. Irjala H, Alanen K, Grenman R, Heikkila P, JoensuuH, Jalkanen S. Mannose receptor (MR) and commonlymphatic endothelial and vascular endothelial receptor(CLEVER)-1 direct the binding of cancer cells to thelymph vessel endothelium. Cancer Res 2003;63:4671–4676.

111. Coussens LM, Werb Z. Inflammation and cancer. Nature2002;420:860–867.

112. Uchida D, Begum NM, Almofti A, et al. Possible role ofstromal-cell-derived factor-1/CXCR4 signaling on lymphnode metastasis of oral squamous cell carcinoma. ExpCell Res 2003;290:289–302.

113. Almofti A, Uchida D, Begum NM, et al. The clinicopa-thological significance of the expression of CXCR4 pro-tein in oral squamous cell carcinoma. Int J Oncol 2004;25:65–71.

114. Samara GJ, Lawrence DM, Chiarelli CJ, et al. CXCR4-mediated adhesion and MMP-9 secretion in head andneck squamous cell carcinoma. Cancer Lett 2004;214:231–241.

115. Wang J, Xi L, Hunt JL, et al. Expression pattern ofchemokine receptor 6 (CCR6) and CCR7 in squamouscell carcinoma of the head and neck identifies a novelmetastatic phenotype. Cancer Res 2004;64:1861–1866.

116. Wang J, Zhang X, Thomas SM, et al. Chemokine re-ceptor 7 activates phosphoinositide-3 kinase-mediatedinvasive and prosurvival pathways in head and neckcancer cells independent of EGFR. Oncogene 2005;24:5897–5904.

117. De Benedetti A, Graff JR. eIF-4E expression and itsrole in malignancies and metastases. Oncogene 2004;23:3189–3199.

118. Estilo CL, O-charoenrat P, Ngai I, et al. The role ofnovel oncogenes squamous cell carcinoma-related onco-gene and phosphatidylinositol 3-kinase p110a in squa-mous cell carcinoma of the oral tongue. Clin CancerRes 2003;9:2300–2306.

119. Sticht C, Hofele C, Flechtenmacher C, et al. Amplifica-tion of cyclin L1 is associated with lymph node metas-tases in head and neck squamous cell carcinoma(HNSCC). Br J Cancer 2005;92:770–774.

120. el-Naggar AK, Dinh M, Tucker S, et al. Genotypicanalysis of primary head and neck squamous carci-noma by combined fluorescence in situ hybridizationand DNA flow cytometry. Am J Clin Pathol 1996;105:102–108.

121. Welkoborsky H-J, Mann WJ, Gluckman JL, Freije JE.Comparison of quantative DNA measurements andcytomorphology in squamous cell carcinomas of theupper aerodigesive tract with and without lymph nodemetastases. Ann Otol Rhinol Laryngol 1993;102:52–57.

122. Sarker SK, Ghufoor K, Patel KS, Tolley NS, ColemanDV. Nuclear DNA content using computerized imagecytometry of squamous cell carcinomas of the head andneck. J Laryngol Otol 1997;111:43–47.

123. Tytor M, Olofsson J. Prognostic factors in oral cavitycarcinomas. Acta Otolaryngol Suppl (Stockh) 1992;492:75–78.

124. Hemmer J, Thein T, Van Heerden WF. The value ofDNA flow cytometry in predicting the development oflymph node metastasis and survival in patients withlocally recurrent oral squamous cell carcinoma. Cancer1997;79:2309–2313.

125. King WWK, Lam PK, Li AKC. DNA ploidy as a predic-tor of cervical metastasis in advanced squamous carci-noma of the tongue. Acta Otolaryngol (Stockh)1995;115:455–458.

126. Byers RM, el-Naggar AK, Lee Y-Y, et al. Can we detector predict the presence of occult nodal metastases inpatients with squamous carcinoma of the oral tongue?Head Neck 1998;20:138–144.

127. Nagata M, Fujita H, Ida H, et al. Identification ofpotential biomarkers of lymph node metastasis in oralsquamous cell carcinoma by cDNA microarray analysis.Int J Cancer 2003;106:683–689.

128. Schmalbach CE, Chepeha DB, Giordano TJ, et al. Mo-lecular profiling and the identification of genes associ-ated with metastatic oral cavity/pharynx squamous cellcarcinoma. Arch Otolaryngol Head Neck Surg 2004;130:295–302.

129. Warner GC, Reis PP, Jurisica I, et al. Molecular classifi-cation of oral cancer by cDNA microarrays identifiesoverexpressed genes correlated with nodal metastasis.Int J Cancer 2004;110:857–868.

130. Cromer A, Carles A, Millon R, et al. Identification ofgenes associated with tumorigenesis and metastaticpotential of hypopharyngeal cancer by microarray anal-ysis. Oncogene 2004;23:2484–2498.


131. Chung CH, Parker JS, Karaca G, et al. Molecular clas-sification of head and neck squamous cell carcinomasusing patterns of gene expression. Cancer Cell 2004;5:489–500.

132. Roepman P, Wessels LF, Kettelarij N, et al. An expres-sion profile for diagnosis of lymph node metastasesfrom primary head and neck squamous cell carcinomas.Nat Genet 2005;37:182–186.

133. Roepman P, Kemmeren P, Wessels LF, Slootweg PJ,Holstege FC. Multiple robust signatures for detectinglymph node metastasis in head and neck cancer. Can-cer Res 2006;66:2361–2366.

134. Bockmuhl U, Schluns K, Schmidt S, Matthias S,Petersen I. Chromosomal alterations during metastasisformation of head and neck squamous cell carcinoma.Genes Chromosomes Cancer 2002;33:29–35.

135. Chen YJ, Lin SC, Kao T, et al. Genome-wide profilingof oral squamous cell carcinoma. J Pathol 2004;204:326–332.

136. Takes RP, Baatenburg de Jong RJ, Keuning J,Hermans J, Schuuring E, van Krieken JHJM. Proteinexpression of cancer associated genes: biopsy materialcompared to resection material. Anticancer Res 1998;18:4787–4792.

137. Ramaswamy S, Ross KN, Lander ES, Golub TR. A mo-lecular signature of metastasis in primary solid tumors.Nat Genet 2003;33:49–54.

138. Takes RP, Baatenburg de Jong RJ, Schuuring E, Litvi-nov SV, Hermans J, van Krieken JHJM. Differences inexpression of oncogenes and tumor suppressor genes indifferent sites of head and neck squamous cell cancer.Anticancer Res 1998;18:4793–4800.

139. Rodrigo JP, Suarez C, Gonzalez MV, et al. Variability ofgenetic alterations in different sites of head and neckcancer. Laryngoscope 2001;111:1297–1301.

140. Freier K, Joos S, Flechtenmacher C, et al. Tissuemicroarray analysis reveals site-specific prevalence of

oncogene amplifications in head and neck squamouscell carcinoma. Cancer Res 2003;63:1179–1182.

141. Ferlito A, Devaney KO, Robbins KT, Rinaldo A. Recom-mendations for studies on detection of neck disease.Oral Oncol 2004;40:967–970.

142. Devaney KO, Rinaldo A, Ferlito A. Micrometastases incervical lymph nodes from patients with squamous car-cinoma of the head and neck—should they be activelysought? Maybe. Am J Otolaryngol 2007;28:271–274.

143. Rinaldo A, Devaney KO, Ferlito A. Immunohistochemi-cal studies in the identification of lymph node microme-tastases in patients with squamous cell carcinoma ofthe head and neck. ORL J Otorhinolaryngol Relat Spec2004;66:38–41.

144. Brennan JA, Mao L, Hruban RH, et al. Molecularassessment of histopathological staging in squamous-cell carcinoma of the head and neck. N Engl J Med1995;332:429–435.

145. Hamakawa H, Fukizumi M, Bao Y, et al. Genetic diag-nosis of micrometastasis based on SCC antigen mRNAin cervical lymph nodes of head and neck cancer. ClinExp Metastasis 1999;17:593–599.

146. Nieuwenhuis EJ, Leemans CR, Kummer A, Denkers F,Snow GB, Brakenhoff RH. Assessment and clinical sig-nificance of micrometastases in lymph nodes of headand neck cancer patients detected by E48 (Ly-6D)quantitative reverse transcription-polymerase chainreaction. Lab Invest 2003;83:1233–1240.

147. Becker MT, Shores CG, Yu KK, Yarbrough WG. Molec-ular assay to detect metastatic head and neck squa-mous cell carcinoma. Arch Otolaryngol Head Neck Surg2004;130:21–27.

148. Michiels S, Koscielny S, Hill C. Prediction of canceroutcome with microarrays: a multiple random valida-tion strategy. Lancet 2005;365:488–492.

149. Roepman P, Holstege FC. Tumor profiling turmoil. CellCycle 2005;4:659–660.


Can biomarkers play a role in the decision about treatment of the clinically negative neck in...

Documents

Transcript of Can biomarkers play a role in the decision about treatment of the clinically negative neck in...