Secondary Logo

Journal Logo

Immunohistochemical expression of epidermal growth factor receptor, PTEN, and E-cadherin in laryngeal squamous cell carcinoma

El-Gendi, Saba Mohameda; Kazem, Amani Husseinb

doi: 10.1097/01.XEJ.0000406595.04109.62
ORIGINAL ARTICLES
Free

Background The prognostic stratification of laryngeal squamous cell carcinoma (LSCC) patients is inadequate as patients affected by tumors with similar clinicopathological parameters, and undergoing similar treatment, may differ in prognosis.

The current study was undertaken to assess the immunohistochemical expression of epidermal growth factor receptor (EGFR), E-cadherin, and phosphatase and tensin homolog deleted in chromosome ten (PTEN) in 52 retrospective LSCC cases. We also investigated the prognostic implication of deregulated marker expression in relation to the clinicopathological features of LSCC, especially neck nodal metastasis.

Results Decreased histologic differentiation, supraglottic tumor location, low or negative EGFR, and low or negative E-cadherin expression were significant predictors of lymph node metastasis (LNM) (P<0.001, P=0.011, P<0.001, and P=0.033, respectively), whereas patients’ age, increased T-stage, cartilage invasion, muscle invasion, and differences in PTEN expression were not predictors of LNM. A receiver operator characteristic curve analysis showed that the predictability of EGFR and E-cadherin expression in discriminating LNM-positive cases was good and poor, respectively. Three models were developed using only the variables that revealed significant association with LNM in bivariate analysis and we tested their predictability to discriminate LNM-positive from LNM-negative cases. All models were significantly better than the basic model containing the constant only (P<0.001 for the three models). A receiver operator characteristic analysis was performed, and it revealed that the predictive power of the first (tumor grade and site) and second (tumor grade, site, and EGFR) models ranged from good to excellent, whereas that of the third model (tumor grade, site, EGFR, and E-cadherin) was excellent. The number of correctly diagnosed cases (overall accuracy) increased only in the second compared with the first model, but the third model had an overall accuracy equal to that of the second model.

Conclusion Nodal metastasis can be predicted by combining tumor grade and site in an equation, and if an immunohistochemical marker is to be added to increase the ability to identify patients with clinically negative nodes, our results suggest it to be EGFR. Although E-cadherin did not increase LNM predictability when added, this may be due to the limitation of the sample size. Thus, according to our results, we recommend an evaluation of these promising molecular markers on a larger sample size in future studies.

aDepartment of Pathology, Faculty of Medicine

bDepartment of Pathology, Medical Research Institute, Alexandria University, Egypt

Correspondence to Saba Mohamed El-Gendi, PhD, 29 Fawzy Moaaz Street, Smouha, Alexandria, Egypt Tel: +2 0122 444 62 60; fax: +2 03 590 72 68; e-mail: sabagendi@yahoo.com

Received August 10, 2011

Accepted September 15, 2011

Back to Top | Article Outline

Introduction

Head and neck squamous cell carcinoma (HNSCC) is a heterogeneous disease involving dysregulation of multiple pathways linked to cellular differentiation, cell cycle control, apoptosis, angiogenesis, and metastasis (Kim and Califano, 2004; Rodrigo et al., 2010). Thus, the identification of biomarkers associated with both locoregional failure and lymph node metastasis (LNM) could predict tumor behavior and guide treatment (Muller et al., 2008).

The epidermal growth factor receptor (EGFR) belongs to the Type I receptor tyrosine kinase family (Almadori et al., 2006). It regulates cell growth in response to binding of its ligands (Cohen et al., 2006; Muller et al., 2008). There is provocative evidence that EGFR expression plays a role in predicting the prognosis of laryngeal squamous cell carcinoma (LSCC) (Almadori et al., 1999, 2006). Tyrosine kinase inhibitors (TKI) and monoclonal antibodies to target EGFR were developed for the treatment of HNSCC; however, resistance to TKI treatment has been observed (Muller et al., 2008). Epithelial to mesenchymal transition was suggested to be a determinant of the sensitivity of cancer cells to EGFR inhibition (Thomson et al., 2005). The Epithelial to mesenchymal transition biomarker E-cadherin was used to predict clinical activity in patients with lung cancer who received combined therapy with erlotinib and chemotherapy (Yauch et al., 2005), and restoring E-cadherin expression enhanced the sensitivity to EGFR-targeted therapy (Witta et al., 2006). Currently, the molecular relation between EGFR and E-cadherin is unclear (Muller et al., 2008).

E-cadherin, a transmembrane glycoprotein (Gao et al., 2006), plays important roles not only in cell adhesion and morphogenesis but also in cellular signal transduction in collaboration with EGFR/ERK-mediated and c-Src-mediated pathways. E-cadherin expression in HNSCC tissue specimens has been reported in several studies (Ferlito et al., 2004, 2006; Muller et al., 2008); however, in those studies, it was not examined in association with EGFR.

The phosphatase and tensin homolog deleted in chromosome 10 (PTEN) is a tumor-suppressor gene located on the 10th chromosome (Sullu et al., 2010), which encodes a dual-specific protein-phospholipid phosphatase involved in the regulation of a variety of signal transduction pathways. Mutation or abnormal expression of the PTEN protein occurs commonly in multiple tumors, and correlates significantly with tumorigenesis and progression of different malignancies (Zheng et al., 2003). It has also been shown that PTEN interacts with cell adhesion molecules such as h-catenin and E-cadherin to inhibit migration and proliferation (Fournier et al., 2009).

The aim of this study was to assess the expression of EGFR, E-cadherin, and PTEN in LSCC, and to identify any possible significant association between the expression of the three markers with each other. A second aim was to assess the implication of marker expression in relation to the clinicopathological features of LSCC (tumor differentiation, invasion, and particularly regional nodal metastasis) in a trial to provide new insights into the biology of tumor progression and metastasis of this disease, which might have clinical implications in the management of LSCC.

Back to Top | Article Outline

Material and methods

The present study included 52 cases of LSCC that were submitted to the Pathology Departments of the Alexandria Faculty of Medicine and Medical Research Institute, during the period between January 2009 and July 2010. All cases were submitted as total laryngectomy specimens with bilateral neck dissection. All cases had a single primary tumor and microscopically free surgical margins, and none of the patients had undergone treatment before surgery.

Clinical characteristics were retrieved from the available clinical records and the macroscopic tumor size was recorded from the retrieved pathology reports. The study was approved by Alexandria University, Faculty of Medicine Research Ethics Committee.

Hematoxylin and eosin-stained sections of the primary tumor and regional lymph nodes were reviewed to confirm the diagnosis and to grade and stage the cases. The histological grade was determined according to the degree of differentiation following the WHO criteria (Cardesa et al., 2005) and were graded as well (grade I), moderately (grade II), and poorly (grade III) differentiated. The stage of disease was determined according to the TNM staging system of the International Union Against Cancer (fifth edition) (Fleming et al., 1997).

One paraffin block, with sufficient tumor tissue representative of the primary tumor and including the tumor border and normal adjacent epithelium, was selected for immunohistochemical staining. Primary tumors were examined without the knowledge of the lymph node status.

A correlation was established between the immunohistochemical profile of the tumors and the clinicopathological features including patient’s age, tumor location (supraglottic, glottic, or infraglottic), grade of differentiation (well, moderately, or poorly differentiated), pathological nodal involvement (positive vs negative), and cartilage and muscle invasion.

Back to Top | Article Outline

Immunohistochemistry

The formalin-fixed paraffin-embedded (FFPE) tissue blocks were cut into 5-μm-thick sections and dried on coated glass slides. The sections were deparaffinized with standard xylene and hydrated through graded alcohol into water. For EGFR, enzymatic antigen retrieval was performed using trypsin (preheated to 37°C) by directly pipetting the solution onto the tissue on the slide and incubating it for 15 min in a 37°C incubator. For both PTEN and E-cadherin, antigen retrieval was performed by heating the tissue sections for 20 min in a microwave oven with citrate buffer (pH 6.0). Tissue slides were incubated overnight at 4°C in a humid chamber with the primary antibody. All three primary antibodies were mouse monoclonal supplied by Neomarkers, Lab Vision Corporation, AR, USA. For EGFR, clone111.6 was used at a dilution of 1 : 200; for PTEN, clone17.A was used at a dilution of 1 : 25; and for E-cadherin, clone 36B5 was used at a dilution of 1 : 20. The antigen–antibody reaction was visualized by Thermo Scientific UltraVision LP Detection System (Neomarkers, Lab Vision Corporation, AR, USA). Immunohistochemical reactions were developed with diaminobenzidine and sections counterstained with Harrris hematoxylin as a final step. After staining, the slides were dehydrated through graded alcohol and mounted with a coverslip. All immunostains were processed manually, with appropriate positive and negative controls included for each batch of slides.

The slides were examined separately by both authors without the knowledge of the clinicopathological data. A score for EGFR immunostaining was utilized and included the percentage of stained tumor cells and the intensity of staining. The percentage of stained cells was graded as follows: 0, no stained cells; 1, 1–30%; 2, 31–50%; and 3, >50%. The intensity of staining was graded from 0 (no staining) to 3 (strong) as compared with the normal epithelium. The sum of the two scores stratified tumors into high expressors if the total score was 5 or 6 or as low expressors if less than 5 (Lee et al., 2010). Cases with a final score of 0 were classified as negative.

E-cadherin scoring was performed by semiquantifying the number of positively stained tumor cells, and the expression was either high or low (more or less than 50%) (Eriksen et al., 2004b). Cases were classified as negative only when there was no single tumor cell showing membranous E-cadherin immunostaining in spite of the presence of positive internal and external controls.

PTEN expression was semiquantitatively assessed for the extent and severity of nuclear staining. The extent of staining was classified as 0, if 0–10% of the tumor cells were stained; 1, if 11–25% of the tumor cells were stained; 2, if 26–50% were stained; and 3, if more than 50% were stained. The intensity of staining was graded as 1, if it was light yellow; 2, if it was dark yellow; and 3, if it was brown. The sum of these two classifications was scored as negative (0) if the score was 0–2, mild (1) if the score was 3–4, and severe (2) if the score was 5–6 (Sullu et al., 2010).

Back to Top | Article Outline

Statistical analysis

Data were fed to the computer using the Predictive Analytics Software (PASW Statistics 18). Qualitative data were described using numbers and percentages. Association between categorical variables was tested using the χ2-test. When more than 20% of the cells had an expected count less than 5, correction for χ2 was conducted using Yate’s correction. Quantitative data were described using median, minimum, and maximum and the mean and SD. The Mann–Whitney was used to test quantitative independent variables. Correlations between two quantitative variables were assessed using Spearman’s ρ test.

A logisitic regression technique was adopted to formulate equations for the prediction of LNM. The model developed was assessed using Model χ2 and Nagelkerke’s R2. The significance of the χ2 indicated that the amount of unexplained information was minimal and the model including the covariates was significantly better than the model including the constant only. The Nagelkerke’s R2 reflected the amount of variation explained by the model. The contribution of the individual predictor in the model was assessed using Wald statistics, which was usually used to ascertain whether a variable was a significant predictor to the outcome.

The diagnostic performance of the different parameters to discriminate LN positive from LN negative cases were evaluated using a receiver operating characteristic (ROC) curve analysis. The cutoff point at which the highest diagnostic accuracy was reached was estimated using the Youden Index. The following statistics were calculated: sensitivity, specificity, positive and negative likelihood ratio, and the positive and negative predictive values. Significance test results were quoted as two-tailed probabilities. The significance of the obtained results was judged at the 5% level.

Back to Top | Article Outline

Results

Clinicopathologic features

The current study included 52 cases of LSCC. All cases were male patients. The age of the patients ranged between 37 and 87 years and the mean age was 56.31±10.08. Out of the 52 cases, three cases (5.8%) were located in the subglottic region, 23 cases (44.2%) were glottic, and 26 cases (50%) were supraglottic. Eight cases (15.4%) were grade I SCC, 26 cases (50%) were grade II, 16 cases (30.8%) were grade III, and two cases (3.8%) were basaloid SCC. Cartilage invasion was positive in 20 cases (38.5%) and negative in 32 cases (61.5%). As for T-stage, four tumors (7.7%) were T1, 16 (30.8%) were T2, 9 (17.3%) were T3, and 23 tumors (44.2%) were T4. A significant negative correlation was noted between the T-stage and the patients’ age (P=0.007).

Back to Top | Article Outline

Immunohistochemical results

Epidermal growth factor receptor immunostaining

In the current study, EGFR membranous expression was seen in the normal laryngeal mucosal epithelial cells, mainly in the basal and the spinous layers, and was used as an internal positive control (Fig. 1a). Membranous EGFR expression was observed in 75% of the tumors studied (n=39). High expression (final score 5 or 6) was observed in 22 cases (Fig. 1b) and low expression (final score below 5) in 17 cases (Fig. 1c). The remaining 13 cases were totally negative (final score 0) (Fig. 1d).

Fig. 1

Fig. 1

Back to Top | Article Outline

E-cadherin immunostaining

E-cadherin membranous expression was seen in the basal and the spinous layers of the normal laryngeal mucosal epithelium, in addition to the acinar cells of the seromucinous glands, and all were used as internal positive controls (Fig. 2a).

Fig. 2

Fig. 2

With regard to E-cadherin immunostaining of neoplastic tissue, 22 cases (42.3%) were classified as high expressers (Fig. 2b) and 11 cases (21.2%) as low expressers (Fig. 2c). The remaining 19 cases (36.5%) revealed total absence of E-cadherin positive immunostaining (Fig. 2d).

Back to Top | Article Outline

PTEN immunostaining

PTEN immunostaining was more marked in the spinous than in the basal cell layer of the normal laryngeal mucosa (Fig. 3a). PTEN nuclear expression was strongly positive in 21 cases (40.4%) (Fig. 3b), mildly positive in 17 cases (32.7%) (Fig. 3c), and negative in 14 cases (26.9%) (Fig. 3d).

Fig. 3

Fig. 3

Both cases of basaloid LSCC were supraglottic in location, muscle invasive, negative for cartilage invasion, T-stage 4, LNM positive, EGFR negative, low E-cadherin expressors, and PTEN nuclear expression negative.

EGFR expression in the studied cases exhibited a significant negative correlation with PTEN expression (ρ=−0.332; P=0.016). No correlation was detected between EGFR and E-cadherin expression (ρ=0.119; P=0.400), or between E-cadherin and PTEN expression (ρ=0.168; P=0.235).

The tumor histologic grade exhibited a highly significant negative correlation with EGFR and E-cadherin expression. Thus, a lower expression was detected in poorly differentiated tumors (ρ=−0.421, P<0.001; and ρ=−0.512, P<0.001, respectively). However, PTEN nuclear expression did not correlate significantly with the tumor grade (ρ=−0.079; P=0.486). EGFR and E-cadherin expression correlated significantly with patients’ age; a lower expression of EGFR and a higher expression of E-cadherin were detected in older patients (ρ=−0.425, P=0.002; and ρ=0.356, P=0.010, respectively). However, PTEN expression did not show a significant difference in relation to the patients’ age (ρ=−0.118; P=0.405). Differences in expression between the different T stages were weakly significant for EGFR expression (P=0.043) and not significant for the expression of both E-cadherin and PTEN (P=0.951 and P=0.190, respectively). Conversely, PTEN was the only marker that showed a significant weak and negative association with cartilage invasion (P=0.040), as the expression of both EGFR and E-cadherin did not show significant differences between the tumors with and without cartilage invasion (P=0.264 and P=0.952, respectively). In contrast, both EGFR and PTEN expression showed significant negative associations with muscle invasion (P=0.048 and P=0.026, respectively), whereas E-cadherin expression did not associate with muscle invasion (P=0.091). EGFR and PTEN expression also showed significant associations with a supraglottic tumor location (P<0.001 and P=0.003, respectively). Conversely, E-cadherin expression did not associate significantly with the tumor location (P=0.073).

In bivariate analysis, the clinicopathologic parameters that were significantly associated with LNM are shown in Table 1. Decreased histologic differentiation (P<0.001) and a supraglottic tumor location (P=0.011) were significant predictors of LNM. Significant differences were also observed in the expression of both EGFR and E-cadherin (P<0.001 and P=0.033, respectively) between the tumors with and without nodal metastases. Thus, nodal metastases were more frequent in the cases with a low or negative expression of EGFR and E-cadherin. Patients’ age, increased T-stage, and cartilage and muscle invasion were not significant predictors of nodal metastases, and no significant differences in PTEN expression were noted between LNM-positive and LNM-negative cases (P=0.625).

Table 1

Table 1

As only EGFR and E-cadherin, and not PTEN, expression showed a significant association with LNM, a ROC curve analysis was performed to assess the predictive power of both EGFR and E-cadherin in discriminating LNM-positive from LNM-negative cases. The predictability of EGFR was good (AUC=0.891) (P<0.001, 95% confidence interval (CI)=0.773–0.960), and using the Youden index, a cut-off value of low or negative EGFR expression was determined to show cases that have a higher incidence of nodal metastases. Similarly, E-cadherin predictability was tested and was poor (AUC=0.661) (P=0.035, 95% CI=0.516–0.786), and a cut-off of low or negative E-cadherin expression was determined to predict cases with a positive LN using the Youden index (Fig. 4).

Fig. 4

Fig. 4

Then, three models were developed to discriminate LNM-positive from LNM-negative cases, using only the variables that revealed a significant association with LNM in a bivariate analysis (tumor grade, site, EGFR, and E-cadherin expression). Table 2 shows the predictive power of the different models tested. The first model included the tumor grade and the site, and its equation is P1=eY1/(1+eY1); Y1=−40.566+[20.430* grade 1]+[42.646* grade 2]+[2.926* (site)]. The second model included the tumor grade, the site, and EGFR expression, and its equation is P2=eY2/(1+eY2); Y2=−40.763+[18.195* grade 1]+[41.569* grade 2]+[21.369* (site)]+[3.585* (EGFR)]. The third model included the tumor grade, the site, EGFR, and E-cadherin expression, and its equation is P3=eY3/(1+eY3); Y3=37.464+[18.116* grade 1]+[38.341* grade 2]+[35.370* (site)]+[3.024* (EGFR)]+[−17.287* E-cadherin]; given that grade 1=1 if MD, otherwise=0, grade 2=1 if PD, otherwise=0, EGFR=1 if low or negative, otherwise=0, E-cadherin=1 if low or negative, otherwise=0, site=1 if supraglottic, otherwise=0.

Table 2

Table 2

All three models were significantly better than the basic model containing the constant only: P<0.001 for the three models. The predictive power of the three models was assessed using ROC analysis. The predictive power of the first and the second models ranged from good to excellent: AUC=0.945, 95% CI 0.844–0.988 and AUC=0.975, 95%CI=0.889–0.997, respectively. However, the predictive power of the third model was excellent: AUC=0.983, 95% CI 0.900–0.997 (Fig. 5).

Fig. 5

Fig. 5

With regard to the nodal metastatic status (whether LNM positive or LNM negative), the first model diagnosed 90.38% of the cases correctly (n=47), compared with only 44 cases (84.62%) using EGFR and 38 cases (73.08%) using E-cadherin. In the second model, when EGFR expression was added to tumor grade and site, the number of correctly diagnosed cases increased, and the model achieved an overall accuracy of 98.08% and only a single case was incorrectly diagnosed as having false-positive (FP) LNM. In the third model, when E-cadherin was added to tumor grade, site, and EGFR expression, it did not increase the number of correctly diagnosed cases (overall accuracy).

Back to Top | Article Outline

Discussion

The prognostic stratification of LSCC patients is inadequate as similar patients, affected by tumors with similar clinicopathological parameters and undergoing the same treatment, may differ considerably in the prognosis, probably due to the biological heterogeneity of LSCCs, and this contributes to a lack of consistency in treatment planning (Ferlito et al., 2004). The indications for comprehensive surgical clearance of the neck, for clinically palpable metastatic lymph nodes (cN+), are obvious; however, the indications for elective selective treatment of clinically negative (cN0) neck seem to be less clear (Ferlito et al., 2006).

Thus, the ability to predict cervical nodal metastasis from primary tumor biopsy specimens would provide tremendous advantages for the determination of optimal therapeutic strategies. This raised interest in identifying molecular markers that would enable the selection of patients at risk for LNM, and as invasion and metastasis are complicated multistep processes, it is likely that more than one marker will be needed to assess an individual patient’s risk of nodal metastases (Rodrigo et al., 2007).

The current study was undertaken to assess the expression of EGFR, E-cadherin, and PTEN in LSCC. We aimed at identifying any possible relation between the expression of the three markers, in addition to the implication of marker expression in relation to the clinicopathological features of LSCC (tumor differentiation, T-stage, invasion, and in particular regional nodal metastasis) in a trial to provide new insights into the biology of tumor progression and metastasis of this disease that might have clinical implications in the management of LSCC.

In agreement with others (Eriksen et al., 2004a, 2004b; Bentzen et al., 2005), we report a highly significant negative correlation between EGFR expression and the tumor grade, and in accordance with the study conducted by the Danish Head and Neck Cancer group (Bentzen et al., 2005), EGFR expression exhibited a weak but significant correlation with T-stage. EGFR expression also showed a significant association with supraglottic tumor location, patients’ age, and muscle invasion, but, not with cartilage invasion.

Our study, in accordance with some researchers (Eriksen et al., 2004a; Mittari et al., 2005) and in opposition to others (Rodrigo et al., 2007), demonstrated that E-cadherin expression correlated significantly only with the tumor grade and the patients’ age, and failed to show a significant association with the tumor location, the T-stage, and both cartilage and muscle invasion. As alterations in cellular adhesion have a fundamental role in regulating the invasive tumor phenotype in HNSCC (Kramer et al., 2005), failure of EGFR expression to associate with invasion was previously explained by the fact that, in addition to reduced cell-cell adhesiveness, loss of adhesion of the epithelial cells to the extracellular matrix is also mandatory to promote tumor cell migration and invasion (Schaller et al., 1992).

In this study, in agreement with another study (Chen et al., 1999), PTEN expression did not associate with patients’ age, tumor grade, T-stage, and LNM, but it revealed a significant association with both cartilage and muscle invasion and with supraglottic tumor location. Data in the literature on the association of PTEN with clinicopathologic parameters and LNM are highly controversial. One previous study (Sullu et al., 2010) found no significant relationship between PTEN expression and tumor size, differentiation, depth of invasion, and LNM, and in a study reporting that there was a decrease in PTEN expression in laryngeal tumors, there was no relationship between PTEN expression and tumor localization, tumor size, differentiation, and stage (Guney et al., 2007). Also, our study, in spite of documenting an association between PTEN expression and invasion of both the cartilage and the muscle, revealed no significant association between PTEN and E-cadherin expression, although PTEN has been previously shown to interact with cell adhesion complexes and to stabilize intercellular junctions (Kotelevets et al., 2001). Aberrations in E-cadherin expression or function were also suggested to cause a loss of PTEN expression in cancers, such as those that frequently occur in the breast, where PTEN expression is lost without identifiable mutations in the PTEN gene itself (Sullu et al., 2010). Conversely, expression of PTEN showed a significant negative correlation with the expression of EGFR (P=0.016).

When all clinicopathologic parameters were tested for the prediction of LNM in a bivariate analysis, higher tumor grade, supraglottic tumor location, low or negative EGFR expression, and low or negative E-cadherin expression were significant predictors of nodal metastasis. These findings are in agreement with previous reports (Hirabayashi et al., 1991; Reid et al., 1991) that observed a significant association between tumor grade (degree of histologic differentiation) and LNM. Our finding that supraglottic tumor location showed the highest rates of LNM (76.9% vs. 34.8% for glottic and 0% for subglottic) is also in agreement with others (Almadori et al., 2006). Previous reports demonstrated that neck node metastasis was mainly a ‘supraglottic issue,’ because of the profuse lymphatic network of the supraglottic larynx (Levendag and Vikram, 1987), and attributed the lower rates of LNM in glottic carcinomas to the paucity of lymphatic drainage of the true vocal cords, in all areas other than the posterior commissure, which makes metastasis of early lesions extremely unlikely. Concerning subglottic cancers, the incidence of cervical metastasis was reported to be 20–30%, but that figure is somewhat obscured by the fact that the primary drainage pattern of these lesions is to the less detectable pretracheal and paratracheal nodes (Harrison, 1971).

The sensitivity and specificity of EGFR for the detection of nodal metastasis were better than those of E-cadherin (89.29% and 79.17% for EGFR vs. 78.57% and 66.67% for E-cadherin, respectively), and the number of correctly diagnosed cases (overall accuracy) was more for EGFR compared with E-cadherin (44/52, 84.62% vs. 38/52, 73.08%), with 83.3% positive predictive value (PPV) and 86.4% negative predictive value (NPV) for EGFR and 73.3% PPV and 72.7% NPV for E-cadherin. Thus, EGFR proved to be a better tool for predicting nodal metastasis if a single immunohistochemical marker is to be applied to primary tumor samples.

Tumor grade and site are assessed routinely, efficiently, and easily, without adding cost or effort to the standard diagnostic work-up of FFPE pathologic specimens. Thus, in a trial to establish cost-effective parameters that aid the preoperative assessment of the nodal status, without even adding the expenses of immunohistochemistry, we created a model to test the combined predictive power of both tumor grade and site used simultaneously to discriminate LNM-positive from LNM-negative cases using ROC curve analysis. Surprisingly, the predictability of this model proved to be better than the predictability of the expression of EGFR or E-cadherin when used alone. This model had a predictive accuracy for cN0 status of 100%, as there were no false-negative cases, which is most important for the goal of achieving clinical relevance. However, the model incorrectly diagnosed five cases as FP (which is similar to the FP value of EGFR expression when used alone, but better (lower) than the FP value of E-cadherin expression). The model reached a sensitivity of 100%, a specificity of 79.17%, with PPV=84.8% and NPV=100%.

With regard to the prediction of nodal metastasis, which subsequently affects treatment planning, a FP diagnosis is better than a false-negative diagnosis. Although a FP diagnosis subjects the patient to unnecessary surgical neck clearance, with its relative comorbidity and comortality, in addition to raising the patient’s anxiety and fear, it is better than a false-negative diagnosis, which reduces patients’ anxiety, but carries worse prognosis in terms of patient outcome and survival.

Thus, in a trial to achieve a higher predictability for nodal metastasis, and to reduce the number of FP diagnosed cases, we established two more models (models 2 and 3). In one model (model 2), we tested the combined predictability of tumor grade, site, and EGFR expression, and in the other model (model 3), we tested the predictive power of all four parameters together (tumor grade, site, EGFR, and E-cadherin expression). In model 2, the number of correctly diagnosed cases (overall accuracy) increased as only a single case was incorrectly diagnosed as being FP for LNM, and the model maintained the 100% sensitivity and 100% NPV of model 1 (that included only the tumor grade and site); however, this model (model 2) showed a higher specificity (95.83%) and a higher PPV (96.6%). Conversely, in model 3, adding E-cadherin expression did not increase the predictability of nodal metastasis, as the single FP-diagnosed case was also not diagnosed correctly by this model.

Our findings, similar to other studies (Mansouri et al., 1988; Liu et al., 2003), noted that E-cadherin expression does not prove to be relevant in predicting LNM. Thus, the clinical significance of an altered expression of E-cadherin in the laryngeal carcinoma remains controversial. How this marker is involved in the invasive and metastatic events still needs to be unmasked.

In contrast to other studies (Chung et al., 2004; Roepman et al., 2005) that suggested that the metastatic state can be deciphered from the primary tumor gene expression pattern, which is laborious, complex, and difficult to implement in routine clinical use, we have achieved an overall accuracy in predicting LNM as high as 98.08%, using only one immunohistochemical marker (EGFR expression) combined with two classic clinicopathologic parameters (tumor grade and site). On the basis of these findings, we recommend that if an immunohistochemical marker is to be used in addition to tumor grade and site in an equation to assess LNM, it is EGFR.

These results are even better than those obtained with the most recent techniques of computed tomography, MRI, ultrasonography, PET, and ultrasound-guided fine-needle aspiration biopsy, which have reached a sensitivity of 80–85% in detecting occult metastases (Van den Brekel et al., 1998). These parameters can be assessed easily on standard FFPE pathologic specimens, thereby facilitating its inclusion in the routine surgical pathology diagnostic practice and subsequent treatment planning.

Back to Top | Article Outline

Conclusion

Our results suggest that an abnormal expression of EGFR, E-cadherin, and PTEN occur in LSCC, and that specific patterns of expression of EGFR and E-cadherin are associated with LNM. Combining the EGFR expression status with tumor grade and site and validating them in an equation can increase the ability to identify patients with clinically negative lymph nodes who are at considerable risk for occult metastases. In those cases, the EGFR expression status may play a role in the decision to treat clinically neck node negative patients (cN0) with a neck dissection or with close follow-up. It also raises the issue of the utility of EGFR targeted therapy (putting into consideration other factors such as patient’s age and comorbidity).

Although E-cadherin did not prove to increase LNM predictability when added to the model, this may be due to the limitation of the sample size. Thus, all the established equations in this study need to be validated on a larger sample size in future studies. It is probably the right time for the integration of molecular markers in the prognostic assessment and treatment selection for patients with LSCC. We strongly recommend the evaluation of these promising molecular markers for clinical practice both by a meta-analysis of data present in the literature and by multidisciplinary and multicentric clinical trials.

Back to Top | Article Outline

Acknowledgements

Conflicts of interest

There are no conflicts of interest.

Back to Top | Article Outline

References

Almadori G, Cadoni G, Galli J, Ferrandina G, Scambia G, Exarchakos G, et al. Epidermal growth factor receptor expression in primary laryngeal cancer: an independent prognostic factor of neck node relapse. Int J Cancer. 1999;84:188–191
Almadori G, Bussu F, Paludettii G. Predictive factors of neck metastases in laryngeal squamous cell carcinoma. Towards an integrated clinico-molecular classification. Acta Otorhinol Ital. 2006;26:326–334
Bentzen SM, Atasoy BM, Daley FM, Dische S, Richman PI, Saunders MI, et al. Epidermal growth factor receptor expression in pretreatment biopsies from head and neck squamous cell carcinoma as a predictive factor for a benefit from accelerated radiation therapy in a randomized controlled trial. JClin Oncol. 2005;23:5560–5567
Cardesa A, Gale N, Nadai A, Zidar NBarnes L, Eveson JW, Reichart P, Sidransky D. Squamous cell carcinoma. World Health Organization classification of tumours pathology and genetics head and neck tumours. 2005 Lyon IARC Press:118–123
Chen RW, Avizienyte E, Roth S, Elivo I, Mäkitie AA, Aaltonen LM, et al. PTEN and LKB1 genes in laryngeal tumours. J Med Genet. 1999;36:943–944
Chung CH, Parker JS, Karaca G, Wu J, Funkhouser WK, Moore D, et al. Molecular classification of head and neck squamous cell carcinomas using patterns of gene expression. Cancer Cell. 2004;5:489–500
Cohen G, Mustafi R, Chumsangsri A, Little N, Nathanson J, Cerda S, et al. Epidermal growth factor receptor signaling is up-regulated in human colonic aberrant crypt foci. Cancer Research. 2006;66:5656–5664
Eriksen JG, Steiniche T, Askaa J, Alsner J, Overgaard J. The prognostic value of epidermal growth factor receptor is related to tumor differentiation and the overall treatment time of radiotherapy in squamous cell carcinomas of the head and neck. Int J Radiat Oncol Biol Phys. 2004a;58:561–566
Eriksen JG, Steiniche T, Søgaard H, Overgaard J Expression of integrins and E-cadherin in squamous cell carcinomas of the head and neck. 2004b;112 APMIS:560–568
Ferlito A, Bradley P, Rinaldo A. What is the treatment of choice for T1 squamous cell carcinoma of the larynx? J Laryngol Otol. 2004;118:747–749
Ferlito A, Rinaldo A, Silver CE, Gourin CG, Shah JP, Clayman GL, et al. Elective and therapeutic selective neck dissection. Oral Oncol. 2006;42:14–25
Fleming ID, Cooper JS, Henson DE, Hutter RVP, Kennedy BJ, Murphy GP, et al. American Joint Committee on Cancer (AJCC) Cancer staging manual. 19975th ed New York Lippincott-Raven
Fournier MV, Fata JE, Martin KJ, Yaswen P, Bissell MJ. Interaction of E-cadherin and PTEN regulates morphogenesis and growth arrest in human mammary epithelial cells. Cancer Res. 2009;69:4545–4552
Gao Z, Tretiakova MS, Liu W, Gong C, Farris PD. Association of E-cadherin, matrix metalloproteinases and tissue inhibitors of metalloproteinases with the progression and metastasis of hepatocellular carcinoma. Mod Pathol. 2006;19:533–540
Guney K, Ozbilim G, Derin AT, Cetin S. Expression of PTEN protein in patients with laryngeal squamous cell carcinoma. Auris Nasus Larynx. 2007;34:481–486
Harrison DF. The pathology and management of subglottic cancer. Ann Otol Rhinol Laryngol. 1971;80:6–12
Hirabayashi H, Koshii K, Uno K, Ohgaki Y, Nakasone T, Fujisawa N, et al. Extracapsular spread of squamous cell carcinoma in neck lymph nodes: prognostic factor of laryngeal cancer. Laryngoscope. 1991;101:502–506
Kim MM, Califano JA. Molecular pathology of head-and-neck cancer. Int JCancer. 2004;112:545–553
Kotelevets L, van Hengel J, Bruyneel E, Mareel M, van Roy F, Chastre E, et al. The lipid phosphatase activity of PTEN is critical for stabilizing intercellular junctions and reverting invasiveness. J Cell Biol. 2001;155:1129–1135
Kramer RH, Shen X, Zhou H. Tumor cell invasion and survival in head and neck cancer. Cancer Metastasis Rev. 2005;24:35–45
Lee CH, Hung HW, Hung PH, Shieh YS. Epidermal growth factor receptor regulates β-catenin location, stability and transcriptional activity in oral cancer. Mol Cancer. 2010;9 Art. No. 64
Levendag P, Vikram B. The problem of neck relapse in early stage supraglottic cancer–results of different treatment modalities for the clinically negative neck. Int J Radiat Oncol Biol Phys. 1987;13:1621–1624
Liu M, Lawson G, Delos M, Jamart J, Chatelain B, Remacle M, et al. Prognostic value of cell proliferation markers, tumour suppressor proteins and cell adhesion molecules in primary squamous cell carcinoma of the larynx and hypopharynx. Eur Arch Oto-Rhino-Laryngol. 2003;260:28–34
Mansouri A, Spurr N, Goodfellow PN, Kemler R. Characterization and chromosomal localization of the gene encoding the human cell adhesion molecule uvomorulin. Differentiation. 1988;38:67–71
Mittari E, Charalabopoulos A, Batistatou A, Charalabopoulos K. The role of E-cadherin/catenin complex in laryngeal cancer. Exp Oncol. 2005;27:257–261
Muller S, Su L, Tighiouart M, Saba N, Zhang H, Shin DM, et al. Distinctive E-cadherin and epidermal growth factor receptor expression in metastatic and nonmetastatic head and neck squamous cell carcinoma: predictive and prognostic correlation. Cancer. 2008;113:97–107
Reid AP, Robin PE, Powell J, McConkey CC, Rockley T. Staging carcinoma: its value in cancer of the larynx. J Laryngol Otol. 1991;105:456–458
Rodrigo JP, Dominguez F, Suárez V, Canel M, Secades P, Chiara MD, et al. Focal adhesion kinase and E-cadherin as markers for nodal metastasis in laryngeal cancer. Archives of Otolaryngol. 2007;133:145–150
Rodrigo JP, García-Carracedo D, González MV, Mancebo G, Fresno MF, García-Pedrero J. Podoplanin expression in the development and progression of laryngeal squamous cell carcinomas. Mol Cancer. 2010;9 Art. No. 48
Roepman P, Wessels LF, Kettelarij N, Kemmeren P, Miles AJ, Lijnzaad P, et al. An expression profile for diagnosis of lymph node metastases from primary head and neck squamous cell carcinomas. Nat Genet. 2005;37:182–186
Schaller MD, Borgman CA, Cobb BS, Vines RR, Reynolds AB, Parsons JT, et al. pp125(FAK), a structurally distinctive protein-tyrosine kinase associated with focal adhesions. Proceedings of the National Academy of Sciences of the United States of America. 1992;89:5192–5196
Sullu Y, Gun S, Atmaca S, Karagoz F, Kandemir B. Poor prognostic clinicopathologic features correlate with VEGF expression but not with PTEN expression in squamous cell carcinoma of the larynx. Diagn Pathol. 2010;5 1 Art. No. 35
Thomson S, Buck E, Petti F, Griffin G, Brown E, Ramnarine N, et al. Epithelial to mesenchymal transition is a determinant of sensitivity of non-small-cell lung carcinoma cell lines and xenografts to epidermal growth factor receptor inhibition. Cancer Res. 2005;65:9455–9462
Van den Brekel MW, Castelijns JA, Snow GB. Diagnostic evaluation of the neck. Otolaryngol Clin N Am. 1998;31:601–620
Witta SE, Gemmill RM, Hirsch FR, Coldren CD, Hedman K, Ravdel L, et al. Restoring E-cadherin expression increases sensitivity to epidermal growth factor receptor inhibitors in lung cancer cell lines. Cancer Res. 2006;66:944–950
Yauch RL, Januario T, Eberhard DA, Cavet G, Zhu W, Fu L, et al. Epithelial versus mesenchymal phenotype determines in vitro sensitivity and predicts clinical activity of erlotinib in lung cancer patients. Clin Cancer Res. 2005;11(24 Pt 1):8686–8698
Zheng HC, Li YL, Sun JM, Yang XF, Li XH, Jiang WG, et al. Growth, invasion, metastasis, differentiation, angiogenesis and apoptosis of gastric cancer regulated by expression of PTEN encoding products. World J Gastroenterol. 2003;9:1662–1666
©2011Egyptian Journal of Pathology