Clinical significance of molecular alterations in histologically negative
surgical margins of head and neck cancer patients
Ana Carolina de Carvalho
a, Luiz Paulo Kowalski
b, Antônio Hugo José Fróes Marques Campos
c,
Fernando Augusto Soares
c, André Lopes Carvalho
d, André Luiz Vettore
a,⇑aLaboratory for Molecular Cancer Biology, Department of Biological Sciences, Federal University of São Paulo, Rua Pedro de Toledo, 669, 11oandar, São Paulo, SP 04039-032, Brazil bDepartment of Head and Neck Surgery, Fundação Antônio Prudente, Hospital do Câncer A C Camargo, Rua Prof. Antonio Prudente, 211, São Paulo, SP 01509-010, Brazil cDepartment of Pathology, Fundação Antônio Prudente, Hospital do Câncer A C Camargo, Rua Prof. Antonio Prudente, 109, 1oandar, São Paulo, SP 01509-010, Brazil dDepartment of Head and Neck Surgery, Fundação PIO XII, Hospital de Câncer de Barretos, Av. Antenor Duarte Villela, 1331, Barretos, SP 14784-400, Brazil
a r t i c l e
i n f o
Article history:
Received 20 July 2011
Received in revised form 8 October 2011 Accepted 26 October 2011
Available online 21 November 2011
Keywords:
Gene expression Head and neck neoplasms Surgical margins Local neoplasm recurrence Second primary neoplasms Field cancerization
Matrix metalloproteinase 9 Parathyroid hormone-related protein
s u m m a r y
The development of locoregional recurrence is the main reason for treatment failure in head and neck squamous cell carcinomas (HNSCC) and the remaining of tumor cells in surgical margins is associated with recurrence. Surgical margins are considered negative based on histologic assessment of the pathological specimen. However, this method lacks sensitivity in identifying cells that already started malignant trans-formation but have not yet developed a pathologic phenotype. We investigated the usefulness of assessing the expression ofPTHLH,EPCAM,MMP9,LGALS1andMETfor the detection of molecular alterations in his-tologically negative surgical margins and determine the correlation of these tumor-related alterations with clinical and prognostic parameters. Differential gene expression was determined by quantitative RT–PCR analyses in normal mucosa, HNSCC and negative margin samples. Thirty-eight percent of the his-tologically negative surgical margins examined were margin-positive for overexpression of at least one of the genes evaluated. Moreover,MMP9andPTHLHoverexpression in the surgical margins was associated with the development of second primary tumors (p= 0.002) and lower rates of local control (log rank test p= 0.022;HR= 4.186;p= 0.035), respectively. These findings demonstrate that the overexpression of tumor-related genes in histologically negative surgical margins is a frequent event. The use of qRT–PCR may be an useful tool in detecting actually negative HNSCC surgical margins and the overexpression of specific genes in these margins could be helpful in the identification of patients with a higher risk of devel-oping second primary tumors and local recurrences, thus aiding the surgeon in the delineation of the HNSCC resection extent and helping in the planning of adjuvant therapy.
Ó2011 Elsevier Ltd.
Introduction
Head and neck cancer is the eighth most common cancer world-wide, with approximately 650,000 new cases per year1and more than 90% being squamous cell carcinomas (HNSCC). Despite differ-ent strategies used in the treatmdiffer-ent of HNSCC cases, a significant percentage of advanced patients still have a poor prognosis, with a high percentage of locoregional and distant recurrences.2,3 Note-worthy, the great genetic and biological heterogeneity of HNSCC makes difficult the comprehension of the molecular carcinogenesis process of these tumors and the development of new therapeutic strategies.
The presence of microscopic tumor in the surgical margins is associated with the high rate of HNSCC relapse and shorter overall survival.4Therefore, obtaining negative surgical margins is of
par-amount importance in reducing these recurrence rates. The extent of tumor surgical resection is conventionally determined during surgery by the surgeon’s gross assessment complemented by an intraoperative examination of the margins based in histologic hematoxylin–eosin (H&E) diagnosis of frozen sections. In spite of the high accuracy ratio reported for this histologic examination (>95%)5–8the elevated rate of treatment failure (around 30%) in pa-tients with histologically negative margins raises concern about the sensitivity of this method.9,10
Therefore, the intraoperative diagnosis of histologically nega-tive margins, although extremely accurate, does not eliminate the possibility of the presence of tumor cells contributing to lo-cal relapses. Taking into account that genetic alterations precede phenotypic changes of the epithelium, molecular assessment of surgical margins could constitute a more sensitive approach to detect the presence of malignant transformed cells in histologi-cally negative surgical margins. Attempts using different molecular strategies have already been performed (i.e., TP53
1368-8375Ó2011 Elsevier Ltd. doi:10.1016/j.oraloncology.2011.10.018
⇑Corresponding author. Tel.: +55 11 5539 6151; fax: +55 11 5571 8806.
E-mail address:andre.vettore@gmail.com(A.L. Vettore).
Contents lists available atSciVerse ScienceDirect
Oral Oncology
j o u r n a l h o m e p a g e : w w w . e l s e v i e r . c o m / l o c a t e / o r a l o n c o l o g y
Open access under the Elsevier OA license.
mutation, IHC analysis of elF4 and MMP9, telomerase activity, microsatellite instability and gene-promoter hypermethylation profile).11–22
Recent studies are focused on identifying molecular factors that may provide useful prognostic information and influence the clin-ical management of HNSCC patients. Ferris et al.23 showed that
PTHLHandEPCAMexpression was able to discriminate between
positive and benign lymph nodes with a high accuracy. Moreover, the IHC detection of MMP9 in surgical margins has showed a posi-tive association with the risk of relapse in HNSCC patients.13,14 However, a significant limitation of IHC evaluations is the long time required to carry out the examination, rendering its use dur-ing surgery. Besides, microarray analyses have foundLGALS1and METoverexpressed in HNSCC samples as compared with normal mucosa.24,25
Previous studies have already supported the theory of field cancerization, which refers to the presence of malignant or prema-lignant changes in the entire field of apparently normal tissue adja-cent to the tumor in response to a carcinogen exposition.26Thus, the goal of this study was to evaluate the ability ofPTHLH,EPCAM,
MMP9,LGALS1 and MET expression in identifying tumor-related
alterations in histologically negative surgical margins. In this way, we hypothesize that the expression profile of these genes in histologically negative margins could act as a more sensitive and useful marker for the detection of molecular alterations associated with local disease control in HNSCC patients.
Materials and methods
Patients
This retrospective study involved tissue specimens from 55 HNSCC patients who underwent tumor resection between 2000– 2008 at A.C. Camargo Hospital and Barretos Cancer Hospital. Only patients diagnosed with primary HNSCC, with surgery as primary curative treatment and with surgical margins samples available were included. Histologically negative surgical margins were avail-able from all 55 patients, while primary HNSCC samples were ob-tained from a subset of 23 patients.
Surgical margin samples comprised mucosa with a clinically normal appearance taken at surgery at least 1 cm from the tumor edge and reported as ‘‘histologically negative’’ after routine intra-operative histologic examination. Prior to RNA extraction and after removal of tumor from the surgical specimen, all remaining lateral margins were processed as a pool and hematoxylin–eosin stained slides were re-evaluated by a senior pathologist in order to confirm the diagnosis. Only surgical margins classified as histologically negative and without the presence of premalignant lesions such as dysplasia, were included in the study. Additionally, 25 oral mu-cosa tissue samples from healthy donors undergoing odontological and pre-prosthetic surgeries were used as normal controls. In-formed consent was obtained from each individual prior to tissue collection and the study protocol was approved by A.C. Camargo Hospital and Barretos Cancer Hospital Ethic Committees and per-formed in accordance with the ethical guidelines of the 1975 Dec-laration of Helsinki.
Gene expression quantification by Real-time RT–PCR (RT–qPCR)
Total RNA was isolated from approximately 0.5 cm3of frozen tissue samples using Trizol reagent (Invitrogen, Frederick, MD) and 2
l
g were reverse transcribed using SuperScripIII First Strand Synthesis Kit (Invitrogen, Frederick, MD) according to manufac-turer’s recommendations. Each cDNA mixture was diluted 10-fold before use.PCR amplification was performed in an ABI 7500 Fast Sequence Detection System using the TaqMan system (Applied Biosystems, Foster City, CA). Primers and probes were obtained from the Inven-toried TaqMan Gene Expression Assays (Applied Biosystems, Foster City, CA). PCR reaction conditions can be provided if requested.
The relative gene expression level was calculated using the mathematical equation 2 (DDCt).27Each sample data was normal-ized on the basis of the average expression of two reference genes (HPRT1andRPLPO) selected by a NormFinder analysis28(data not shown). The average expression value obtained from 10 normal oral mucosa samples was used as calibrator. Additional 15 normal oral mucosa samples were used to calculate specificity levels. Final results were expressed asn-fold differences in gene expression be-tween the evaluated sample and the calibrator sample.
Statistical analysis
Sensitivity and specificity of each individual gene in distin-guishing HNSCC samples from normal controls were calculated. A Receiver Operating Characteristic (ROC) curve was constructed with the expression level of each gene in the HNSCC samples and in 15 normal controls. The area under the receiver operator charac-teristic (AUC) identified optimal sensitivity and specificity levels to distinguish HNSCC patients from healthy individuals, allowing the definition of cut-off values for each gene.
The association between the clinical variables and molecular data was evaluated using Chi-square or Fisher’s exact tests. The ef-fects of the molecular markers and clinical variables on overall, lo-cal recurrence-free and locoregional recurrence-free survival, and potential field lesions (local, regional and second primary tumors in the UADT) were estimated by Kaplan–Meier method and com-pared by the log-rank test. Cox proportional hazard regression analysis was performed to estimate the hazard ratios (HR) for clin-ical and molecular variables. All statistclin-ical analysis was performed using the statistical software SPSS 17.0 for Windows and a level of significance ofp< 0.05 was adopted.
Results
Follow-up and characteristics of the patients
Clinical and pathological data of the 55 HNSCC patients enrolled in this study are presented inTable 1. The mean follow-up was 30.6 months (median = 20.9 months; range 0.43–98.13 months). Mean interval from diagnosis to recurrence was 12.2 months (range 5–24 months). Treatment failure was observed in 34.5% of the cases (19/55): 11 patients (20.0%) showed local recurrence, 15 (27.3%) locoregional recurrence and 4 (7.3%) patients developed second primary tumors (SPT) in the upper aerodigestive tract (lung, tongue, esophagus and lip) defined according to the criteria proposed by Warren and Gates.29Therefore, 19 (34.5%) patients developed potential field lesions (local or regional recurrences and SPT in the UADT).
Molecular analysis of primary tumors and surgical margins
The expression levels of 5 genes,PTHLH,EPCAM,MMP9,LGALS1
andMET, were assessed by qRT-PCR assays.
Table 1
Clinical and pathological data of the patients enrolled in the study.
Characteristic Number of cases (%) MMP9 EPCAM PTHLH
Negative Positive p Negative Positive p Negative Positive p
Total patients 55 (100.0) Age
Mean, median, range 58.25, 59.0, 28–85 years
<60 years 31 (56.4) 26 (83.9) 5 (16.1) 0.136 28 (90.3) 3 (9.7) 0.534 30 (96.8) 1 (3.2) 0.107 >60 years 24 (43.6) 16 (66.7) 8 (33.3) 21 (87.5) 3 (12.5) 20 (83.3) 4 (16.7)
Tobacco consumption
No 12 (21.8) 8 (66.7) 4 (33.3) 0.296 11 (91.7) 1 (8.3) 0.609 11 (91.7) 1 (8.3)
Yes 43 (78.2) 34 (79.1) 9 (20.9) 38 (88.4) 5 (11.6) 39 (90.7) 4 (9.3) 0.702
Alcohol consumption
No 18 (32.7) 13 (72.2) 5 (27.8) 0.426 16 (88.90 2 (11.1) 0.649 17 (94.4) 1 (5.6) 0.467
Yes 37 (67.3) 29 (78.4) 8 (21.6) 33 (89.2) 4 (10.8) 33 (89.2) 4 (10.8)
Clinical nodal status
N- 24 (43.6) 17 (70.8) 7 (29.2) 0.272 20 (83.3) 4 (16.7) 0.234 22 (91.7) 2 (8.3) 0.607
N+ 30 (54.5) 25 (83.3) 5 (16.7) 28 (93.3) 2 (6.7) 27 (90.0) 3 (10.0)
Unknown 1 (1.9)
Clinical tumor status
T1–T2 26 (47.2) 21 (80.8) 5 (19.2) 0.61 25 (96.2) 1 (3.8) 0.114 25 (96.2) 1 (3.8) 0.199
T3–T4 28 (50.9) 21 (75.0) 7 (15.0) 23 (82.1) 5 (17.9) 24 (85.7) 4 (14.3)
Unknown 1 (1.8)
Clinical TNM stage*
Initial (I/II) 16 (29.1) 11 (68.8) 5 (31.3) 0.245 15 (93.8) 1 (6.3) 0.418 15 (93.8) 1 (6.3) 0.532 Advanced (III/IV) 38 (69.1) 31 (81.6) 7 (18.4) 33 (86.8) 5 (13.2) 34 (89.5) 4 (10.5)
Not available 1 (1.8)
Primary tumor site
Oropharynx 6 (11.0) 5 (83.3) 1 (16.7) 0.583 5 (83.3) 1 (16.7) 0.023 6 (100) 0 (0) 0.201
Oral cavity 41 (74.5) 32 (78) 9 (22) 39 (95.1) 2 (4.9) 38 (92.7) 3 (7.3)
Larinx 8 (14.5) 5 (62.5) 3 (37.5) 5 (62.5) 3 (37.5) 6 (75) 2 (25)
Treatment
Surgery + radiotherapy 35 (65.6) 27 (77.1) 8 (22.9) 0.553 30 (85.7) 5 (14.3) 0.28 31 (88.6) 4 911.4) 0.394 Surgery only 20 (36.4) 15 (75.0) 5 (25.0) 19 (95.0) 1 (5.0) 19 (95.0) 1 (5.0)
Vascular embolization
Negative 43 (78.2) 33 (76.7) 10 (23.3) 0.529 38 (88.4) 5 (11.6) 0.646 41 (95.3) 2 (4.7) 0.052
Positive 11 (20) 8 972.7) 3 (27.3) 10 (90.9) 1 (9.1) 8 (72.7) 3 (27.3)
Unknown 3 (5.4)
Perineural invasion
No 30 (54.5) 25 (83.3) 5 (16.7) 0.2 26 (86.7) 4 (13.3) 0.494 29 (96.7) 1 (3.3) 0.095
Present 22 (40) 15 (68.2) 7 (31.8) 20 (90.9) 2 (9.1) 18 (81.8) 4 (18.2)
Not available 3 (5.5)
Second primary tumor
No 51 (92.7) 42 (82.4) 9 (17.6) 0.002 45 (88.2) 6 (11.8) 0.621 46 (90.2) 5 (9.8) 0.675
Yes 4 (7.3) 0 (0) 4 (100) 4 (100) 0 (0) 4 (100) 0 (0)
*Clinical stage according toTNM Classification of Malignant Tumours – 6th ed.
A.C.
de
Carvalho
et
al.
/Oral
Oncology
48
(2012)
Table 2
Capability of the gene expression to discriminate HNSCC and historically free surgical margins samples from healthy oral mucosa.
Gene AUC Cut-off HNSCC samples Healthy oral mucosa Surgical margins
Sensitivity#% (95% CI) Specificityà
% (95% CI) Expression frequency§% (95% CI)
MMP9 1 10 91 (71.9–98.9) 100 (78.2–100.0) 23.6 (13.23–37.02)
PTHLH 0.87 5 74 (51.6–89.8) 93 (68.0–99.8) 10.9 (4.11–22.45)
EPCAM 0.6 10 30 (13.2–52.9) 100 (78.2–100.0) 9.1 (3.02–19.95)
LGALS1 0.83 3 83 (61.2–95.5) 73 (44.9–92.2) NE
MET 0.84 3 48 (26.8–69.4) 86 (59.5–98.3) NE
Cut-off value adopted to consider a positive gene expression. # Evaluated in 23 tumor samples.
àEvaluated in 15 healthy oral mucosa.
§Evaluated in 55 surgical margin samples, NE = not evaluated.
the normal samples, trying to adjust these values to the highest specificity and sensitivity rates and maximize the capacity of dis-tinction between HNSCC and healthy subjects (Fig. S1). Hence,
for MMP9 and EPCAMa 10X cut-off was adopted, for PTHLH a
cut-off of 5X whilefor MET and LGALS1 a 3X cut-off was used (Table 2). To be considered overexpressed, the gene expression level should be higher than the cut-off value established for each gene.
MMP9(91%),LGALS1(83%),PTHLH(74%),MET(48%) andEPCAM (30%) were frequently found overexpressed in the 23 HNSCC sam-ples (high sensitivity).MMP9,PTHLHandEPCAMwere rarely over-expressed in the 15 healthy oral mucosa samples (0%; 7% and 0%, respectively), confirming their overexpression as cancer-associated (high specificity). On the other hand,METandLGALS1were found overexpressed in 14% and 27%, respectively of the normal controls, suggesting a lack of specificity (Fig. 1;Table 2), and were excluded from further analyses.
Therefore, based on the high specificity and sensitivity presented
byMMP9,EPCAMandPTHLH, we further investigated the expression
of these genes in 55 histologically negative surgical margins. Even though all surgical margins examined were diagnosed as histologi-cally negative, 36.4% (20/55; 95%CI= 23.81–50.44) showed overex-pression of at least one of these 3 genes.MMP9was overexpressed in
Figure 2Receiver operating characteristic (ROC) curve obtained forMMP9 over-expression in histologically negative surgical margins according to development of second primary tumors.
Table 3
Results of univariate analysis of selected prognostic factors for overall survival.
Characteristic Number
of cases
Number of deaths
5-year overall survival
Pvalue (log-rank)
Hazard ratio for death (95% CI)
Pvalue (Cox)
Tobacco consumption
No 12 2 76.2 Reference
Yes 53 22 31.7 0.073 3.469 (0.814–14.773) 0.092
Alcohol consumption
No 18 7 37.5 Reference
Yes 37 17 42.3 0.599 1.266 (0.524–3.061) 0.6
Clinical nodal status
N 24 7 53.8 Reference
N+ 30 17 26 0.04 2.488 (1.016–6.096) 0.046
Unknown 1
Clinical tumor status
T1–T2 26 6 69.8 Reference
T3–T4 28 18 19.6 0.009 3.249 (1.284–8.224) 0.013
Unknown 1
Clinical TNM stage*
Initial (I/II) 16 2 81.7 Reference
Advanced (III/IV) 38 22 25.4 0.009 5.543 (1.297–23.687) 0.021
Unknown 1
Primary tumor site
Oropharynx 6 5 0 Reference
Oral cavity 41 15 51.3 0.268 (0.102–0.802) 0.017
Larynx 8 4 25 0.041 0.428 (0.112–1.635) 0.215
Treatment
Surgery + radiotherapy 35 19 34 Reference
Surgery only 20 5 52.4 0.231 0.551 (0.204–1.483) 0.238
Vascular embolization
Negative 43 15 46.3 Reference
Positive 11 8 27.3 0.033 2.474 (1.044–5.862) 0.04
Unknown 3
Perineural invasion
Negative 30 8 48.1 Reference
Positive 22 14 30.2 0.02 2.690 (1.126–6.424) 0.026
Unknown 2
MMP9 overexpression in surgical margins
Negative 42 17 39.3 Reference
Positive 13 7 32.7 0.293 1.6 (0.6661–3.871) 0.297
EPCAM overexpression in surgical margins
Negative 49 21 41.5 Reference
Positive 6 2 0 0.439 1.615 (0.474–5.501) 0.443
PTHLH overexpression in surgical margins
Negative 50 21 40 Reference
Positive 5 3 25 0.407 1.664 (0.493–5.619) 0.412
23.6% (13/55) of the surgical margins tested,EPCAM in10.9% (6/55)
andPTHLHin 9.1% (5/55) (Fig. 1;Table 2).
Expression profile and clinical data correlations
The expression profile of the 3 selected genes (MMP9,EPCAM
andPTHLH) in histologically negative surgical margins was
ana-lyzed for potential correlations with patients’ clinicopathological parameters.
There was no association between the overexpression ofPTHLH
andEPCAMin the histologically negative surgical margins and
rele-vant patient features (Table 1). Besides that,MMP9overexpression in the surgical margins was correlated with the development of SPT (p= 0.002) (Table 1) with high sensitivity, specificity and accu-racy, 100% (4/4; 95% CI= 39.77–100.00), 82.3% (42/51; 95% CI= 69.13–91.59) and 83.6% (46/51; 95%CI= 71.19–92.23), respec-tively. ROC curve analysis showed an area under the curve of 0.912
(95%CI= 82.9–99.4) (Fig. 2), evidencingMMP9as a potential marker for the diagnosis of patients at risk for developing SPT.
Nineteen (34.5%) of the HNSCC patients included in this study died of cancer and the 5-year overall survival was 46.0%. Kaplan– Meier and Cox proportional hazard regression analyses were per-formed to estimate the effects of conventional prognostic factors and molecular markers on overall and local recurrence-free sur-vival (Table 3andTable 4), locoregional recurrence-free survival and potential field lesions (Table S1 and Table S2).
As expected, well known clinical prognostic factors as clinically positive lymph nodes (p= 0.04), T3–T4 tumors (p= 0.009), advanced tumors (p= 0.009), vascular embolization (p= 0.033), and perineu-ral invasion (p= 0.02) were significantly associated with reduced OS (Table 3). No association was found between expression of the investigated genes in the surgical margins and patients’ OS (Table 3). T3–T4 tumors (p= 0.024), perineural invasion (p= 0.002) andPTHLH overexpression (p = 0.022) were significantly associated with reduced local recurrence-free survival (Fig. 3; Table 4). Patients
Table 4
Results of univariate analysis of selected prognostic factors for local recurrence-free survival.
Characteristic Number of cases
Number of local recurrences
5-Year local recurrence-free survival
Pvalue (log-rank) Hazard Ratio for local recurrence (95% CI)
Pvalue (Cox)
Tobacco consumption
No 12 1 90 Reference
Yes 43 10 67.3 0.251 3.133 (0.400–24.533) 0.277
Unknown 1 0
Alcohol consumption
No 18 4 69.2 Reference
Yes 37 7 74.3 0.985 1.012 (0.296–3.459) 0.985
Unknown 1 0
Clinical nodal status
N- 24 4 78.4 Reference
N+ 30 7 68.2 0.503 1.521 (0.442–5.231) 0.506
Unknown 1 0
Clinical tumor status
T1–T2 26 2 87.7 Reference
T3–T4 28 9 58.9 0.024 4.959 (1.067–23.042) 0.041
Unknown 1 0
Clinical TNM stage*
Initial (I/II) 16 1 88.9 Reference
Advanced (III/IV) 38 10 66.6 0.094 4.905 (0.626–38.434) 0.13
Unknown 1 0
Primary tumor site
Oropharynx 6 1 75 Reference
Oral cavity 41 8 73.1 0.790 (0.098–6.350) 0.825
Larynx 8 2 68.6 0.783 1.357 (0.122–15.032) 0.804
Treatment
Surgery + radiotherapy 35 10 68 Reference
Surgery only 20 1 88.9 0.13 0.233 (0.030–1.823) 0.233
Vascular embolization
Negative 43 7 78.6 Reference
Positive 11 4 50 0.149 2.404 (0.702–8.230) 0.162
Unknown 3 0
Perineural invasion
Negative 30 2 88.5 Reference
Positive 22 9 47.1 0.002 7.976 (1.720–36.983) 0.008
Unknown 5 0
MMP9 overexpression in surgical margins
Negative 46 9 71.5 Reference
Positive 14 2 78.8 0.783 0.806 (0.174–3.741) 0.783
EPCAM overexpression in surgical margins
Negative 49 9 74.1 Reference
Positive 6 2 62.5 0.329 2.111 (0.455–9.798) 0.34
PTHLH overexpression in surgical margins
Negative 50 8 78.3 Reference
Positive 5 3 25 0.022 4.186 (1.104–15.877) 0.035
overexpressing PTHLH were found to have a significant worse prognosis and had a significantly higher risk of developing local recurrence (HR= 4.186; 95%CI= 1.104–15.877;p= 0.035).
The multivariate Cox regression analysis showed significant dif-ference in overall survival and local recurrence-free survival by clinical features well established as poor prognostic factors: clini-cal tumor status (p= 0.003 andp= 0.003) and perineural invasion (p= 0.006 and p= 0.014) (Table 5). For locoregional recurrence, perineural invasion (p= 0.012) was significantly relevant and vas-cular embolization (p= 0.008) and perineural invasion (p= 0.033) were associated with potential field lesions (Table S3). In this anal-ysis, no significant associations were observed between clinical features and overexpression of the genes tested, probably due to the small sample size analyzed in the study.
Discussion
Surgical margin status is described as an important prognostic factor associated with high risk of local relapses and decrease in survival rates for HNSCC patients.4,30For this reason, the histologic assessment of HNSCC surgical margins by the pathologist at the time of surgery plays an important role in the success of patient treatment. Unfortunately, local recurrence occurs in around 20% of the patients with histologically negative margins.4,8–10 One explanation for this resides in the fact that intraoperative histo-logic evaluation of the mucosa cannot detect molecular alterations
that do not involve phenotypic changes in the cells but set them up in the tumorigenesis track.
In this study we selected 5 genes to have their expression eval-uated in HNSCC, histologically normal surgical margins and healthy oral mucosa samples. All genes selected have been previ-ously found specifically overexpressed in HNSCC. However, no study has been conducted, as of now, to evaluate them as potential markers for detection of molecular changes in histologically nega-tive surgical margins and as putanega-tive predictors of local failure in HNSCC cases.
Although all patients included in this study showed histologi-cally negative margins, 20% (11/55) presented local relapse, sug-gesting that molecular changes not detected by standard microscopic analysis, could be a cause of malignant transformation and the poor outcome. This finding is in accordance with the hypothesis introduced by Slaughter et al.26 about a residual ‘‘al-tered field’’ in the area adjacent to the tumor that could be the leading cause of treatment failure due to a high propensity to de-velop local recurrences and second primary tumors in the head and neck after treatment.
According to our results, 36.4% of the histologically negative margins harbored overexpression of at least one of the 3 selected genes,MMP9,EPCAM, andPTHLH.
EPCAM (epithelial cell adhesion molecule)encodes an antigen associated with carcinoma that is located in the cell membrane and acts as a cell adhesion molecule. Andratschke et al.32 demon-Figure 3Kaplan–Meier curve comparing the probability of local recurrence-free survival in patients with positive or negative overexpression ofPTHLHin histologically negative surgical margins (5-years local recurrence-free survival: 79.3% vs. 25.0%, respectively; p = 0.022).
Table 5
Results of multivariate analysis of selected prognostic factors for local recurrence-free survival and overall survival.
Characteristic Hazard ratio for local recurrence (95% CI)
Pvalue (Cox) Hazard ratio for death (95% CI)
Pvalue (Cox)
Clinical tumor status
T3–T4 12.058 (2.324–62.561) 0.003 4.834 (1.691–13.817) 0.003
Perineural invasion
strated, by conventional RT–PCR, thatEPCAMcould be aberrantly expressed in the HNSCC surgical margins, but the correlation with clinical data was not assessed. In the present study, no significant association was observed between the expression of this gene in the histologically negative surgical margins and relevant patient features.
PTHLH (parathyroid hormone-like hormone), also known as PTHrP, regulates endochondral bone development and epithelial-mesenchymal interactions during the formation of mammary glands and teeth. In breast cancer, the increased expression of PTHLH receptors is associated with proliferation, invasion and in-crease of intracellular cyclic AMP levels.33,34In HNSCC, this gene acts as a mediator of epidermal growth factor receptor (EGFR) and seems to be associated with cell growth and invasive activ-ity.35 However, Dunne et al.36 failed to find any correlation be-tween the presence of PTHLH and the pattern of tumor invasion. Our results show that the overexpression of this gene in the histo-logically negative surgical margins was significantly correlated with reduced local recurrence-free survival and an increased risk to develop local recurrences.
Another common reason for treatment failure in HNSCC is the development of second primary tumors (SPT).37 Until recently, it was thought that development of SPT represented progression of multiple separate genetically altered mucosal foci. However, re-cent studies have been reporting that at least a proportion of these SPT arise from residual portions of a single contiguous pre-neoplastic field after the complete resection of the index tumor. According to them, a stem cell acquires genetic alterations and forms a patch with genetically altered daughter cells. As a result of subsequent genetic alterations, the stem cell escapes normal growth control, gains growth advantage, and develops into an expanding clone. The lesion laterally displaces the normal epithe-lium and additional genetic hits give rise to various subclones within the field. Different clones diverge at a certain point with respect to genetic alterations but do share a common clonal ori-gin, and as a result of the process of clonal divergence and selec-tion, eventually a subclone evolves into invasive cancer.39 Our results suggest that one of these genetic alterations could be the aberrant expression ofmetalloproteinase-9(MMP9) gene. Dif-ferent studies have already reported the overexpression of this protein in HNSCC tumor samples40,41and a significant correlation with regional recurrence and distant metastasis.42Jordan et al.43 demonstrated that, after a 7-year follow-up, several cases of dys-plasia with higher MMP9 expression progressed to HNSCC. We foundMMP9overexpressed in 23.6% of the histologically negative surgical margins and this was associated with the development of SPT (p= 0.002). AUC analyses also showed that this marker has a high accuracy as a diagnostic test for evaluation of the risk for SPT development.MMP9encodes a gelatinase that degrades type IV collagen, the major constituent of basement membrane. The lateral spread of clones from malignant tumors involves the occurrence of multiple factors necessary for cell motility to pen-etrate the extracellular matrix.44 Thus, the involvement of this protein in the process of transformation of precancerous lesions in invasive tumors and its role in the degradation of extracellular matrix may explain the results of the association betweenMMP9 overexpression and the development of SPT.
In summary, our results showed thatMMP9,EPCAM, andPTHLH are frequently and specifically overexpressed in histologically neg-ative margins of resected HNSCC. In spite of the small number of samples evaluated, we demonstrated that the overexpression of
PTHLHandMMP9in histologically negative HNSCC margins is
sig-nificantly correlated to a high risk of local failure and development of SPT. for the first time, we report a significant association be-tween the expression ofMMP9in histologically negative surgical margins of HNSCC cases and the development of SPT and that
the overexpression of PTHLH in histologically negative margins contributes to a reduced local recurrence-free survival. Based in these results, we may speculate that the expression of these genes in HNSCC surgical margins, diagnosed as tumor-free by conven-tional histopatologic analyses, could be a helpful biomarker to identify subjects at risk of new neoplastic evolution. Further vali-dation of these results requires studies with larger patient groups and longer follow-up period, but by achieving a good predictive negative value, this qRT–PCR approach could constitute an alterna-tive tool to guide the surgeon in the delineation of the HNSCC resection extent, in the adoption of a more intensive surveillance during follow-up and help in the planning of adjuvant therapy. Fi-nally, it must be stressed that histopathological examination re-mains irreplaceable in clinical practice, allowing the detection of histologically negative surgical margins during surgery (by H&E examination). Hence, if qRT–PCR analysis of surgical margins is to be used in clinical practice, it should take part in a multimodal-ity diagnostic protocol.
Conflict of interest statement
All authors declare that they have no conflicts of interests.
Disclosure
The authors declare that they have no competing interests.
Acknowledgements
A.C.C. is recipient of scholarship from Fundação de Amparo à Pesquisa do Estado de São Paulo (FAPESP; grant number 2007/ 56245-0). A.L.C. has a National Counsel of Technological and Scien-tific Development scholarship (CNPq). A.L.V. has a National Counsel of Technological and Scientific Development scholarship (CNPq; grant number 302360/2008-5) and a Fundação de Amparo à Pesqu-isa do Estado de São Paulo scholarship (FAPESP; grant number 2008/58460-9).
Appendix A. Supplementary data
Supplementary data associated with this article can be found, in the online version, atdoi:10.1016/j.oraloncology.2011.10.018.
References
1. Parkin DM, Bray F, Ferlay J, Pisani P. Global cancer statistics, 2002.CA Cancer J Clin2005;55(2):74–108.
2. Hinerman RW, Mendenhall WM, Morris CG, Amdur RJ, Werning JW, Villaret DB. Postoperative irradiation for squamous cell carcinoma of the oral cavity: 35-year experience.Head Neck2004;26(11):984–94.
3. Leemans CR, Tiwari R, Nauta JJ. Van dW, I, Snow GB. Recurrence at the primary site in head and neck cancer and the significance of neck lymph node metastases as a prognostic factor.Cancer1994;73(1):187–90.
4. Looser KG, Shah JP, Strong EW. The significance of ‘‘positive’’ margins in surgically resected epidermoid carcinomas.Head Neck Surg1978;1(2):107–11. 5. Ikemura K, Ohya R. The accuracy and usefulness of frozen-section diagnosis.
Head Neck1990;12(4):298–302.
6. Gandour-Edwards RF, Donald PJ, Wiese DA. Accuracy of intraoperative frozen section diagnosis in head and neck surgery: experience at a university medical center.Head Neck1993;15(1):33–8.
7. Gandour-Edwards RF, Donald PJ, Lie JT. Clinical utility of intraoperative frozen section diagnosis in head and neck surgery: a quality assurance perspective.
Head Neck1993;15(5):373–6.
8. Spiro RH, Guillamondegui Jr O, Paulino AF, Huvos AG. Pattern of invasion and margin assessment in patients with oral tongue cancer. Head Neck
1999;21(5):408–13.
9. Parsons JT, Mendenhall WM, Stringer SP, Cassisi NJ, Million RR. An analysis of factors influencing the outcome of postoperative irradiation for squamous cell carcinoma of the oral cavity.Int J Radiat Oncol Biol Phys1997;39(1):137–48. 10. Laccourreye O, Hans S, Menard M, Garcia D, Brasnu D, Holsinger FC. Transoral
II. An analysis of the incidence, related variables, and consequences of local recurrence.Arch Otolaryngol Head Neck Surg2005;131(7):592–9.
11. van Houten VM, Leemans CR, Kummer JA, Dijkstra J, Kuik DJ, van den Brekel MW, et al. Molecular diagnosis of surgical margins and local recurrence in head and neck cancer patients: a prospective study.Clin Cancer Res2004;10(11): 3614–20.
12. Poeta ML, Manola J, Goldenberg D, Forastiere A, Califano JA, Ridge JA, et al. The Ligamp TP53 assay for detection of minimal residual disease in head and neck squamous cell carcinoma surgical margins. Clin Cancer Res 2009;15(24): 7658–65.
13. Nathan CO, Franklin S, Abreo FW, Nassar R, de Benedetti A, Glass J. Analysis of surgical margins with the molecular marker eIF4E: a prognostic factor in patients with head and neck cancer.J Clin Oncol1999;17(9):2909–14. 14. Nathan CO, Amirghahri N, Rice C, Abreo FW, Shi R, Stucker FJ. Molecular
analysis of surgical margins in head and neck squamous cell carcinoma patients.Laryngoscope2002;112(12):2129–40.
15. Goldenberg D, Harden S, Masayesva BG, Ha P, Benoit N, Westra WH, et al. Intraoperative molecular margin analysis in head and neck cancer. Arch Otolaryngol Head Neck Surg2004;130(1):39–44.
16. Martone T, Gillio-Tos A, De ML, Fiano V, Maule M, Cavalot A, et al. Association between hypermethylated tumor and paired surgical margins in head and neck squamous cell carcinomas.Clin Cancer Res2007;13(17):5089–94.
17. Brennan JA, Mao L, Hruban RH, Boyle JO, Eby YJ, Koch WM, et al. Molecular assessment of histopathological staging in squamous-cell carcinoma of the head and neck.N Engl J Med1995;332(7):429–35.
18. Partridge M, Li SR, Pateromichelakis S, Francis R, Phillips E, Huang XH, et al. Detection of minimal residual cancer to investigate why oral tumors recur despite seemingly adequate treatment.Clin Cancer Res2000;6(7):2718–25. 19. Fabricius EM, Gurr U, Wildner GP. Telomerase activity levels in the surgical
margin and tumour distant tissue of the squamous cell carcinoma of the head-and-neck.Anal Cell Pathol2002;24(1):25–39.
20. Ries JC, Hassfurther E, Steininger H, Kloss FR, Wiltfang J, Girod SC, et al. Correlation of telomerase activity, clinical prognosis and therapy in oral carcinogenesis.Anticancer Res2001;21(2A):1057–63.
21. Temam S, Casiraghi O, Lahaye JB, Bosq J, Zhou X, Julieron M, et al. Tetranucleotide microsatellite instability in surgical margins for prediction of local recurrence of head and neck squamous cell carcinoma.Clin Cancer Res
2004;10(12 Pt 1):4022–8.
22. Xu L, Chow J, Bonacum J, Eisenberger C, Ahrendt SA, Spafford M, et al. Microsatellite instability at AAAG repeat sequences in respiratory tract cancers.
Int J Cancer2001;91(2):200–4.
23. Ferris RL, Xi L, Raja S, Hunt JL, Wang J, Gooding WE, et al. Molecular staging of cervical lymph nodes in squamous cell carcinoma of the head and neck.Cancer Res2005;65(6):2147–56.
24. Ginos MA, Page GP, Michalowicz BS, Patel KJ, Volker SE, Pambuccian SE, et al. Identification of a gene expression signature associated with recurrent disease in squamous cell carcinoma of the head and neck. Cancer Res
2004;64(1):55–63.
25. Cromer A, Carles A, Millon R, Ganguli G, Chalmel F, Lemaire F, et al. Identification of genes associated with tumorigenesis and metastatic potential of hypopharyngeal cancer by microarray analysis. Oncogene
2004;23(14):2484–98.
26. Slaughter DP, Southwick HW, Smejkal W. Field cancerization in oral stratified squamous epithelium; clinical implications of multicentric origin. Cancer
1953;6(5):963–8.
27. Livak KJ, Schmittgen TD. Analysis of relative gene expression data using real-time quantitative PCR and the 2(-Delta Delta C(T)) Method. Methods
2001;25(4):402–8.
28. Andersen CL, Jensen JL, Orntoft TF. Normalization of real-time quantitative reverse transcription-PCR data: a model-based variance estimation approach to identify genes suited for normalization, applied to bladder and colon cancer data sets.Cancer Res2004;64(15):5245–50.
29. Warren S, Gates O. Multiple primary malignant tumors. A survey of the literature and a statistical study.Am. J. Cancer1932;16:1358–414.
30. Haque R, Contreras R, McNicoll MP, Eckberg EC, Petitti DB. Surgical margins and survival after head and neck cancer surgery. BMC Ear Nose Throat Disord
2006;6:2.
32. Andratschke M, Hagedorn H, Luebbers CW, Schmitt B, Lang S, Zeidler R, et al. Limited suitability of EpCAM for molecular staging of tumor borders in head and neck cancer.Anticancer Res2006;26(1):153–8.
33. Southby J, Kissin MW, Danks JA, Hayman JA, Moseley JM, Henderson MA, et al. Immunohistochemical localization of parathyroid hormone-related protein in human breast cancer.Cancer Res1990;50(23):7710–6.
34. Luparello C, Birch MA, Gallagher JA, Burtis WJ. Clonal heterogeneity of the growth and invasive response of a human breast carcinoma cell line to parathyroid hormone-related peptide fragments. Carcinogenesis 1997;18(1): 23–9.
35. Yamada T, Tsuda M, Ohba Y, Kawaguchi H, Totsuka Y, Shindoh M. PTHrP promotes malignancy of human oral cancer cell downstream of the EGFR signaling.Biochem Biophys Res Commun2008;368(3):575–81.
36. Dunne FP, Bowden SJ, Brown JS, Ratcliffe WA, Browne RM. Parathyroid hormone related protein in oral squamous cell carcinomas invading the mandible.J Clin Pathol1995;48(4):300–3.
37. Sturgis EM, Miller RH. Second primary malignancies in the head and neck cancer patient.Ann Otol Rhinol Laryngol1995;104(12):946–54.
39. Braakhuis BJ, Tabor MP, Kummer JA, Leemans CR, Brakenhoff RH. A genetic explanation of Slaughter’s concept of field cancerization: evidence and clinical implications.Cancer Res2003;63(8):1727–30.
40. Chambers AF, Matrisian LM. Changing views of the role of matrix metalloproteinases in metastasis.J Natl Cancer Inst1997;89(17):1260–70. 41. Gray ST, Wilkins RJ, Yun K. Interstitial collagenase gene expression in oral
squamous cell carcinoma.Am J Pathol1992;141(2):301–6.
42. Katayama A, Bandoh N, Kishibe K, Takahara M, Ogino T, Nonaka S, et al. Expressions of matrix metalloproteinases in early-stage oral squamous cell carcinoma as predictive indicators for tumor metastases and prognosis.Clin Cancer Res2004;10(2):634–40.
43. Jordan RC, Macabeo-Ong M, Shiboski CH, Dekker N, Ginzinger DG, Wong DT, et al. Overexpression of matrix metalloproteinase-1 and -9 mRNA is associated with progression of oral dysplasia to cancer.Clin Cancer Res2004;10(19):6460–5. 44. Stetler-Stevenson WG, Aznavoorian S, Liotta LA. Tumor cell interactions with
the extracellular matrix during invasion and metastasis.Annu Rev Cell Biol
