RESEARCH ARTICLE

Relationship of immunohistochemical biomarker expressionand lymph node involvement in patients undergoing surgicaltreatment of NSCLC with long-term follow-up

Ana María Gómez & Jose Ramón Jarabo Sarceda & Jose Antonio L. García-Asenjo &

Cristina Fernandez & Susana Hernandez & Julian Sanz & Elena Fernandez &

Joaquin Calatayud & Antonio Torres & Florentino Hernando

Received: 22 September 2013 /Accepted: 23 December 2013# International Society of Oncology and BioMarkers (ISOBM) 2014

Abstract We try to identify the relationship between immuno-histochemical marker expression and lymph node involvementin a cohort of 282 patients followed for 5 years after curativeresection for NSCLC. In 189 patients (67 %), lymph nodes wereunaffected while 93 patients (33 %) showed nodal involvement.The expression of 15 molecular markers was determined fromeach patient by tissue-array immunohistochemistry. Univariateanalysis indicated significantly higher expression of E-cadherin,γ-catenin, p27, and p53 in patients with lymph node

involvement. In those with unaffected nodes, p16 and Rb wereexpressed. E-cadherin expressionwas related to a 50%mortalityreduction in patients with node involvement (hazard ratio (HR)0.5; p=0.017). c-erbB-2 expression was correlated witha 3.4-fold increase in mortality compared to patientswithout expression of this marker in subjects without nodeinvolvement (HR 3.41; p=0.017). Multivariate analysis identi-fied c-erbB-2 (HR 2.22; p=0.089) and p27 (HR 1.44; p=0.019)as prognostics of mortality while Rb (HR 0.74) indicated agood prognosis. The expression of proteins encoded by onco-genes and tumor suppressor genes was different according tolymph node involvement. The increased mortality relatedto c-erbB-2 expression in patients with unaffected lymphnodes would suggests a need for adjuvant treatment.

Keywords Non-small cell lung cancer . Gene expression .

Lymph nodes . Tissue array . Prognosis . c-erb-2

Introduction

Non-small cell lung carcinoma entails a high mortality, evenamong patients undergoing resection with curative purpose;recent studies have related the expression of certain oncogenesand tumor suppressor genes to the definition of patients withdifferent prognosis [1–3]. Some authors such as Brambillaet al. [4] even suggest the use of those markers for an earlydetection of the disease or its relapse. In an effort to identifypatients with a worse prognosis who might benefit from anintensive follow-up or even from adjuvant treatment follow-ing surgical resection, this study was designed to find markersthat could be correlated with lymph node involvement, whosedetection could help for the development of individualizedtreatment strategies.

A. M. Gómez : J. R. Jarabo Sarceda (*) : E. Fernandez :F. HernandoThoracic Surgery Department, Hospital Clínico San Carlos, Institutode Investigación Sanitaria del Hospital Clínico San Carlos (IdISSC),C/Martin Lagos s/n, 28040 Madrid, Spaine-mail: 3sfvideo@gmail.com

J. A. L. García-AsenjoDepartment of Pathology, Hospital Universitario Príncipe deAsturias, Alcalá de Henares, Madrid, Spain

C. FernandezDepartment of Epidemiology, Hospital Clínico San Carlos, Institutode Investigación Sanitaria del Hospital Clínico San Carlos (IdISSC),Madrid, Spain

S. Hernandez : J. SanzDepartment of Pathology, Hospital Clínico San Carlos, Instituto deInvestigación Sanitaria del Hospital Clínico San Carlos (IdISSC),Madrid, Spain

J. CalatayudThoracic Surgery Department, Hospital Clínico San Carlos, Institutode Investigación Sanitaria del Hospital Clínico San Carlos (IdISSC),Madrid, Spain

A. TorresGeneral Surgery Department, Hospital Clínico San Carlos, Institutode Investigación Sanitaria del Hospital Clínico San Carlos (IdISSC),C/Martin Lagos s/n, 28040 Madrid, Spain

Tumor Biol.DOI 10.1007/s13277-013-1599-9

In a large tumor series, tissue-microarrays (TMA) technol-ogy is effective for the study of DNA expression profiles byfluorescent in situ hybridization, RNA expression by in situhybridization, or protein expression profiles by immunohisto-chemistry (IHC). Thanks to this technology, hundreds oftumor markers can be analyzed using a minimum amount oftissue. The entire tumor series, positive and negative controlsand the technique’s internal controls are simultaneously ana-lyzed avoiding biases. Interrelations or synergistic effects ofmarkers determine a need for multivariate analysis including asufficient number of cases to provide meaningful information.

Our main objective was to identify molecular markers thatmight indicate a better or worse prognosis through lymphnode involvement and survival in patients with resected lungcancer and long follow-up by analyzing the immunohisto-chemical profile performed with TMA. This informationwould allow us to optimize not only postoperative surveil-lance programs but also the indication for adjuvant therapies.

Materials and methods

A retrospective study was conducted on a target cohort ofpatients in which data had been prospectively collected. In allthe patients, a complete surgical resection of I to IIIA stagednon-small cell lung cancer had been performed over the periodbetween 1999 and 2005. Complete resection included freeresection margins, mediastinal lymph node dissection, lack ofextracapsular nodal involvement, and the most distant lymphnode removed free of tumor [5, 6]. Routine preoperative stagingincluded clinical examination, chest radiography, and computedtomography scan (CT) of the chest and upper abdomen.Positron emission tomography (PET) was not available at thattime. When mediastinal lymph node involvement wassuspected by radiological criteria mediastinoscopy was per-formed. Systematic mediastinal lymph node dissection wasperformed in all patients. Exclusion criteria were: Stages IIIBand IV, preoperative radiotherapy or chemotherapy and historyof previous malignancy. Inclusion criteria included patientswith non-small cell lung cancer undergoing surgery with cura-tive purpose, survival of at least 4 weeks after surgery, over 5-years follow-up and adequate tumor biopsy that provided ade-quate sample for IHC analysis. Consent was obtained frompatients in order to use their tumor samples for this study.

Table 1 illustrates a flow chart of the patient selection proce-dure [7]. Characteristics of our cohort in terms of demographic,pathologic, and surgical parameters are shown in Table 2. Mostof them are similar to those reported in other studies [8–10].

Clinical data of resected patients are prospectively included ina database, weekly updated thanks to a follow-up programmonitored by the thoracic surgical team. This follow-up scheduleincludes visits every 3 months during the first 2 years aftersurgery and every 6months thereafter up to 5 years or evenmore.

According to intraoperative findings, participants wereassigned to group 1 if no lymph nodes were involved or group2 if at least one lymph node was involved (N1 or N2). Tumorsamples were histologically assessed (including tissue-arrayreading) by two independent pathologists, unknowing the clin-ical data of the patients [11].

Battery of suppressor genes and oncogenes related to cellproliferation and adhesion was examined [12]. Proteinsencoded by tumor suppressor genes such as retinoblastomaprotein (Rb), p16, p27, and p53 [13–22] act as cell cycleinhibitors. The retinoblastoma gene (Rb-1) codes for a proteinthat plays a role in cell division. The Rb protein is down-regulated by protein p16, which is encoded by a gene with itssame name. As a cyclin-dependent kinase (CDK) inhibitor,p27 binds to cyclin E/Cdk2 complex and negatively regulatescell proliferation. Protein p53 induces transcription of severalgenes inducing cell cycle arrest and apoptosis.

Further proteins involved in the cell cycle were also ana-lyzed, such as phosphorylating enzymes CDK 1 and CDK 2and their activating subunits, cyclins A and E, all of theminvolved in proliferating intranuclear pathways [23–26].

Tyrosine kinase receptors were also analyzed [27].Mutations in the proto-oncogenes that code for these receptorsentail structural modifications leading to a decrease in theirresponse to external signals. The oncogene c-erbB-2 [28] (alsoknown as HER-2/neu) codes for the membrane protein kinasep185 [29–31]. Its expression triggers the activation of cellularproliferation signals even in the absence of the specific growthfactor ligand. The proto-oncogene c-Kit [32] codes for atransmembrane receptor with a role in hematopoiesis, game-togenesis, and melanogenesis. It is also closely related to thegenesis of different cancers. The epidermal growth factorreceptor (EGFR), when induced by specific ligands, triggersa cascade of intracellular signals leading to an increase in cellactivity [33].

Expression of myc proteins, involved in the regulation ofcellular proliferation, differentiation, and apoptosis, were alsoinvestigated [34, 35].

Cellular adhesion proteins keep tissular integrity. One ofthe first steps in tumor invasion is loss of adhesion, favoringtumoral cells to get into the blood vessels and move to otherorgans. Among this kind of proteins, β-catenin, γ-catenin,and E-cadherin were selected for our study. Besides its struc-tural role in adhesion junctions, β-catenin is also a transcrip-tional coactivator of genes involved in cell proliferation[36–41]. Finally, other proteins also contribute to tumor inva-sion, such as cathepsin D [42], are also included.

Tumor samples embedded in paraffin were histologicallyexamined to select two suitable areas in each tumor for thedevelopment of TMA. Areas of inflammation, necrosis, fibro-sis or adjacent tissues were avoided. The expression of 15molecular markers was assessed in the 282 tumor samples.Non-tumor samples embedded in paraffin taken from the

Tumor Biol.

same patients served as controls. In addition, each of thesamples was incorporated in the recipient block in duplicate

to ensure the representativity of the sample versus tumorheterogeneity.

Antigen unmasking was conducted in a pressure cham-ber and immunohistochemical staining was evaluated in anautomatic immunostainer. Tissue-array sections were sub-mitted to the Centro Nacional de InvestigacionesOncológicas for immunohistochemical labeling. The anti-bodies, antigens, and dilutions used were as follows: c-erbB-2/neu (DAKO 0485, 1:250); p53 (DAKO DO7,1:40); EGFR (Novocastra, 1:25); C-myc (Novocastra,1:250); E-cadherin (DAKO 36135, 1:50); Rb (BD Phar,1:250); p27 (Transduction Lab, 1:1000); p16 (Santa Cruz,1:50) and γ-catenin (DAKO 1:25); cathepsin D (Dako. RefM7243); CDK1 (BD Transduction Labs. Ref 610038);CDK2 (Neomarkers Clon 8D4. Ref MS-463-P1); cyclin A(Novocastra. Clon 6E6. Ref NCL-Ciclina A); cyclin E(Novocastra. Clon 13-A-3. Ref NCL-Ciclina E); and c-Kit(Dako. Ref A4502).

Immunohistochemical staining was scored by two inde-pendent pathologists in a double-blinded fashion. Sampleswere scored for each antibody according to the intensity ofstaining and number of immunoreactive cells and coded as

Table 1 Patient selection procedure. Period: 1999–2005. N=562 patients

previous or concurrent tumor: non 475 previous or concurrent tumor: yes 87

prior chemotherapy or radiotherapy: non 422 prior chemotherapy or radiotherapy: yes:53

88sey:VIrobIIIetatS433non:VIrobIIIetatS

complete tumor resection: yes 316 complete tumor resection: non 18

Postoperative mortality: non 305 Postoperative mortality: yes 11

5 years of follow up: yes 285 5 years of follow up: non 20

biopsy material available: yes 282 biopsy material available non 3

Table 2 Frequency of distributions of clinical and pathological factors(N=282 patients)

Number Percent (%)

Sex

Male 267 94.7

Female 15 5.32

Age (year)

≤65 138 48.9

>65 144 51.0

Histological type

Squamous cell ca 174 61.7

Other 108 38.2

Nodal involvement

Group 1 N0 189 67.0

Group 2 N1 36 12.7

N2 57 20.2

Tumor Biol.

negative or positive according to date available in publishedstudies [28, 43, 44].

Interrelations or synergistic effects among these markersdetermine a need for multivariate analysis including a suffi-cient number of cases and representative markers in order toprovide meaningful information.

The following variables were considered in the statisticalanalysis: lymph node involvement (yes/no), age (≥65 or<65 years), gender, and tumor histology (squamous cell/other).

For univariate analysis of qualitative variables we used thePearson Chi-squared test. Fisher’s exact test was used forvariables with two categories. When more than two categorieswere compared, the ratio of similitude test was used.

Overall survival for all causes was calculated by theKaplan-Meier method. Curves were compared using the log-rank test. In the multivariate analysis, we used Cox’s propor-tional hazards model to assess correlations with clinico-pathological variables and determine expression patternsfound to be associated with survival in the univariate analysis.Data were stratified by lymph node involvement, and hazardratios (HR) are provided along with 95 % confidence interval(95%CI). p<0.05was considered as significant. All statisticaltests were performed using IBM SSPS 19 software (IBMCompany, Chicago, USA).

Results

As shown in Table 2, our cohort consisted of 282 patients: 267(95 %) men and 15 (5 %) women, with mean age of 65 years(range 41–87). A total of 189 (67 %) of the patients had nolymph node involvement (group 1) and 93 (33 %) had lymphnode involvement (group 2), 36 of them having interlobar(N1) and 57 with mediastinal (N2) lymph node involvement.Median number of resected nodes was 19 (range 11–61).There were 174 (62 %) cases of squamous cell carcinomaand 108 (38 %) cases of adenocarcinoma. The most frequentsurgical technique was lobectomy (51 %).

Follow-up (median 90 months) was completed in all pa-tients since it was one of the inclusion criteria. Tumor recur-rence occurred in 145 patients (51 %), and 187 patients (66 %)died during follow-up. Notably, in 17 patients, the cause ofdeath was not directly related to tumor, being heart disease andrespiratory disease the main reasons.

Protein expression

In the univariate analysis (Table 3), we found higher expressionof E-cadherin (82.8 %, p<0.001) γ-catenin (69.6 %, p<0.001),p27 (48.4 %, p=0.011), and p53 (69.6 %, p=0.026) in speci-mens from patients with lymph node involvement. Those with-out affected lymph nodes showed overexpression of p16(53.4%, p=0.003) and pRb (76.2%, p=0.040). A trend towards

more c-myc expression was also found in this subgroup(p=0.054).

Survival

For the whole series, expression of p27 (Fig. 1a) was associ-ated with a lower 3-year survival rate (47 % versus 60 %;p=0.016). Those exhibiting pRb reached 3-year survival ratesof 59 versus 41% among patients with low expression of pRb(p=0.016) (Fig. 1b). Overexpression of γ-catenin (Fig. 1c)was also related to lower survival (58 versus 45 %), beingthese differences close to significance (p=0.06). Finally, pa-tients without c-erbB-2 expression (Fig. 1d) showed a 3-yearsurvival rate of 57 versus 40 % among patients with high c-erB-2 immunostaining (p=0.06).

When comparing results depending on lymph node in-volvement, we also found differences. Thus, E-cadherin ex-pression (Fig. 2a) was correlated with a reduction of 50 %(hazard ratio=0.5) in the mortality rate in subjects with affect-ed lymph nodes (95 % confident interval 0.28–0.88;p=0.017). Furthermore, c-erbB-2 expression (Fig. 2b) wasassociated with a 3.4-fold increase in mortality (hazard ratio=3.41) in patients with unaffected lymph nodes (95 % confidentinterval 1.25–9.36; p=0.05).

Expression of E-cadherin, p27, p53, c-myc, Rb, c-erbB-2(Fig. 3), p16, and the clinical variables as age, histologicalsubtype, and lymph node involvement were included in theCox regression multivariate model (Table 4).

Adjusting for age, histological tumor type and lymph nodeinvolvement, the expression of c-erbB-2 and p27 was associ-ated with higher mortality rates (HR 2.22 95 % CI 0.89–5.56;p=0.089 and 1.44 95% CI 1.06–1.95; p=0.019, respectively).

Table 3 Univariate analysis: markers expression in patients according tolymph node involvement

Marker N0 (%) N1 or N2 (%) pValue

E-Cadherin 0 48.1 17.2 <0.0011 51.9 82.8

γ-Cadherin 0 60.8 30.1 <0.0011 39.2 69.9

p27 0 67.2 51.6 0.0111 32.8 48.4

p53 0 43.9 30.1 0.0261 56.1 69.9

p16 0 46.6 65.6 0.0031 53.4 34.4

Rb 0 23.8 35.5 0.0401 76.2 64.5

c-myc 0 37.6 48.4 0.0541 62.4 51.6

cerB2 0 97.9 98.9 0.551 2.1 1.1

Tumor Biol.

Overexpression of pRb was associated with good prognosis(HR 0.74 95 % CI 0.53–1.03). No correlations were detectedfor the remaining eight molecular markers. We found nodifferences in the expression of biomarkers according to his-tological subtype (Table 5).

Discussion

This study was designed trying to elucidate whether proteinprofiles in resected non-small-cell lung carcinoma(NSCLC) were different according to lymph node involve-ment. Studies such as those by Takada et al. [45] suggest

that in squamous cell tumors and adenocarcinoma, a panelof genes is able to predict mediastinal node involvementwith an accuracy rate over 95 %. In a study published in2010 including 196 patients undergoing complete resectionfrom 1985 to 1997 in which EGFR, c-erbB-2, c-kit, cyclinB1, and cyclin D1 among others were examined, Grossiet al. [1] concluded that by combining several clinicalfactors with molecular markers, identification of a subsetof patients with stage IIIAN2 NSCLC with a high risk oftumor recurrence using conventional immunohistochemis-try was possible. In a recent article by Donnem et al., thevascular endothelial growth factor-A was identified as anegative independent prognostic factor in NSCLC and its

Fig. 1 Kaplan-Meier survival curves according to expression of biomolecular markers. a p27, b pRB, c γ-catenin, and d c-erB-2/Her2

Tumor Biol.

prognostic capacity seems to be mainly related to lymphnode involvement [46]. Singhal et al. [12] reviewed theprognostic implications of biological markers and arguedthat the information provided would be useful to anticipatedisease progression and select patients amenable to receivea individualized treatment. Finally, in a meta-analysis basedon data from 462 articles and 12 reviews from May 1987 toOctober 2005, Zhu et al. proposed p16, p27, and β-cateninas good candidates for prognosis assessment [47].

Clinical studies addressing NSCLC have described thatdysfunction of the cadherin/catenin complex is related totumor dedifferentiation, lymphatic invasion, and worse prog-nosis [48]. Sulzer et al. [49] reported significant inverse cor-relation between E-cadherin expression and lymph nodestage. Consistent with this finding, we observed a significantlyhigher expression of E-cadherin (82.8 %, p<0.001) and γ-catenin (69.6 %, p<0.001) in tumor samples from patientswith lymph node metastasis, yet the expression of E-cadherinwas linked to a mortality rate that was relatively reduced by50 % compared to patients lacking the expression of this

marker. In other words, this marker is overexpressed in patientswith lymph node involvement but without effect in survival.Ucvet et al. [50] reported that patients in whom positive stain-ing for E-cadherin was detected showed a higher 5-year sur-vival, as indicated by our univariate analysis, although he couldnot confirm that result in the multivariate model, as we couldnot either. Among this controversy, studies such as that byNakashima et al. [51] have shown that patients with low E-cadherin expression have a shorter survival and those with highimmunostaining live longer. Some authors even consider E-cadherin inactivation as an early event in tumor progression.

In a study by Esposito et al. [43], among patients undergo-ing surgical treatment for NSCLC with long-term follow-up,negative correlation was detected between lymph node in-volvement and p21 or p16 expression suggesting a possible

Fig. 2 Kaplan-Meier survival curves in the subgroup of patients with lymph node involvement according to expression of aE-cadherin and b c-erB-2/Her2

Fig. 3 Positive strong (+++) immunostaining for c-erB-2/Her2 (×200magnification)

Table 4 Multivariate analysis

gl pValue HR 95.0 % CI for HR

Min Max

E-Cadherin 1 0.776 0.955 0.693 1.315

p27 1 0.019 1.438 1.061 1.949

p53 1 0.510 0.902 0.664 1.226

c-myc 1 0.420 0.878 0.641 1.204

Rb 1 0.071 0.740 0.533 1.026

c-erbB-2, 1 0.089 2.220 0.886 5.563

P16 1 0.091 0.984 0.727 1.331

Age 1 0.157 1.013 0.995 1.030

TH 1 0.876 0.976 0.724 1.317

N 1 0.006 0.627 0.448 0.877

Variables included age, histological subtype, e-cadherin, p27, p53, c-myc,pRb, c-erB-2, p16

Tumor Biol.

role of these proteins in disease progression. Mohamed et al.[52] detected p16 predicting good prognosis in N2 patientsand proposed surgical resection as the goal treatment forpatients with p16 expression. We observed significantly great-er expression levels of p16 (53.4 %, p=0.003) in patients withunaffected lymph nodes. However, this finding did not corre-late with a benefit in associate survival. In addition, Sterlacciet al. [53] linked p16 to a poor prognosis in smoker malepatients under 65 years with early staged disease. Espositoet al. [17] when stratifying lung tumor samples according top21 or p16 expression observed that overall survival wasshorter in patients showing both negative p21 and p16 expres-sion. This finding is in line with the hypothesis proposed byKaye indicating that most lung cancer samples have aninactivated RB/p16 tumor suppressor pathway [25]. Worseprognosis has been observed in patients whose tumors showedaberrant p53 and p16 expression [16]. Finally, in a univariatestudy by Hommura et al. [44] examining p16 and pRb expres-sion, no significant differences were detected in terms ofsurvival although the number of patients in his study waslow (n=76). In our study, we noted significantly greater p27expression (48.4 %, p=0.011) in tumors with lymph nodeinvolvement related to a lower survival rate (p=0.016). Thisfinding differs from the results of Hayashi et al. [18], relatinglow levels of p27 with poor prognosis. His sample size withalso low (n=98), Similarly, Sion-Vardy et al. [54] was unableto identify p27 as an independent prognostic factor in hisstudy among 92 cases.

In addition, we observed significantly greater pRb expres-sion in patients without lymph node metastasis related to ahigher survival rate (p=0.016). These results are comparableto those of Caputi et al. [14]. In 2002, he first showed the roleof lack of pRb expression as a predictor of mortality amongpatients with cancer. In our series, pRb expression got a 70 %decreasing of mortality rates.

In our study, univariate analysis revealed that p53 expres-sion was higher in patients with lymph node involvement butwithout effect on survival. In line with this, Cheng et al. [55]correlated p53 with a poor prognosis among patients withstages I and II NSCLC. Once more, Esposito et al. [56] alsofound that p53 was a marker of bad prognosis in adenocarci-nomas. Kawasaki et al. [19] reached the same conclusion butfor advanced tumors.

High levels of p185 (c-erbB-2 product) have been detectedin patients with bad prognosis and high recurrence rates [29].Shi et al. [31] were able to correlate p185 expression withlymph node metastasis in squamous cell carcinomas but not inadenocarcinomas. Au et al. also found a worse prognosisamong patients with adenocarcinoma and increased expres-sion of p185 [57]. Grob et al. [28] detected its lack of expres-sion in patients with no lymph node involvement. In our study,we noted that patients with low c-erbB-2 expression showed ahigher value (p=0.06). Moreover, mortality was 3.4-foldhigher in patients without affected nodes who expressed c-erbB-2, suggesting the need for adjuvant treatment in thatselected group of patients.

Conclusion

Among patients undergoing surgical treatment of NSCLCwith long-term follow-up, the expression of proteins encodedby oncogenes and tumor suppressor genes is different depend-ing on whether mediastinal lymph nodes are involved or not.In our study analyzing expression of 15 proteins related to cellproliferation and cellular adhesion, we found that the presenceof high expression of c-erb2 and p27 significantly increasedmortality rates. In the other hand, pRb expression leads to animportant reduction in mortality. The fact that patients withoutlymph node involvement overexpressing c-erb2 showed a 3.4-fold increase in mortality, could be a reason to propose anykind of adjuvant treatment after surgery. However, the numberof samples without c-erb2 expression is too low for transla-tional conclusions. Immunohistochemistry analysis ofresected samples of NSCLC continues to be a feasible wayof defining subgroups of patients who could benefit frompostoperative intensive surveillance or even adjuvant treat-ments. Further studies investigating this issue could giverelevant information in order to reach a clinical correlationwith these findings.

Table 5 Univariate analysis: markers expression in patients according tohistological subtype (squamous/adenocarcinoma)

Histological subtype Pearson’s chi-squared test

Squamous AD

N % N % p value

E-Cadherin 0 69 39.9 25 30.1 0.1291 104 60.1 58 69.9

γ-Cadherin 0 91 52.6 36 43.4 0.1671 82 47.4 47 56.6

p 27 0 108 62.4 51 61.4 0.8801 65 37.6 32 38.6

p53 0 64 37.0 35 42.2 0.4261 109 63.0 48 57.8

p 16 0 93 53.8 41 49.4 0.5131 80 46.2 42 50.6

Rb 0 40 23.1 28 33.7 0.0721 133 76.9 55 66.3

c-myc 0 71 41.0 34 41.0 0.9911 102 59.0 49 59.0

c-erB 2 0 171 98.8 80 96.4 0.1831 2 1.2 3 3.6

AD adenocarcinoma, 0 negative expression, 1 positive expression

Tumor Biol.

Funding source Fundación Mutua Madrileña.

Conflicts of interest None

References

1. Grossi F, Spizzo R, Bordo D, et al. Prognostic stratification of stageIIIA pN2 non-small cell lung cancer by hierarchical clustering anal-ysis of tissue microarray immunostaining data. J Thorac Oncol.2010;5:1354–60.

2. Yokoi S, Yasui K,MoriM, et al. Amplification and overexpression ofskp2 are associated with metastasis of non-small-cell lung cancers tolymph nodes. Am J Pathol. 2004;165:175–80.

3. Moriya Y, Iyoda A, Kasai Y, et al. Prediction of lymph node metas-tasis by gene expression profiling in patients with primary resectedlung cancer. Lung Cancer. 2009;64:86–91.

4. Brambilla C, Fievet F, Jeanmart M, et al. Early detection of lungcancer: role of biomarkers. Eur Respir J. 2003;21(Suppl39):36s–44.

5. Grupo de Trabajo de la SEPAR Normativa actualizada. Sobrediagnóstico y estadificación del carcinoma broncogénico. ArchBronconeumol. 1998;34:437–52.

6. Rami-Porta R, Mateu-Navarro M, Freixinet J, et al. Type of resectionand prognosis in lung cancer. Experience of a multicentre study. Eur JCardiothorac Surg. 2005;28:622–8.

7. Fernández E. Estudios epidemiologicos (STROBE). Med Clin(Barc). 2005;125:43–8.

8. Bria E, Milella M, Sperduti I, et al. A novel clinical prognostic scoreincorporating the number of resected lymph-nodes to predict recur-rence and survival in non-small-cell lung cancer. Lung Cancer.2009;66:365–71.

9. Lee JG, Lee CY, Park IK, et al. Number of metastatic lymph nodes inresected non-small cell lung cancer predicts patient survival. AnnThorac Surg. 2008;85:211–5.

10. Lardinois D, Suter H, Hakki H, et al. Morbidity, survival, and site ofrecurrence after mediastinal lymph-node dissection versus systematicsampling after complete resection for non-small cell lung cancer. AnnThorac Surg. 2005;80:268–75.

11. Bollen E, Van Duin CJ, Theunissen PHMH, et al. Mediastinal lymphnode dissection in resected lung cancer: morbidity and accuracy ofstaging. Ann Thorac Surg. 1993;55:961–6.

12. Singhal S, Vachani A, Antin-Ozerkis D, et al. Prognostic implicationsof cell cycle, apoptosis, and angiogenesis biomarkers in non-smallcell lung cancer: a review. Clin Cancer Res. 2005;11:3974–86.

13. Baldi A, Esposito V, De Luca A, et al. Differential expression of theretinoblastoma gene family members pRb/p105, p107, andpRb2/p130 in lung cancer. Clin Cancer Res. 1996;2:1239–45.

14. Caputi M, Groeger AM, Esposito V, et al. Loss of pRb2/p130expression is associated with unfavorable clinical outcome in lungcancer. Clin Cancer Res. 2002;8:3850–6.

15. Catzavelos C, Tsao MS, DeBoer G, et al. Reduced expression of thecell cycle inhibitor p27Kip1 in non-small cell lung carcinoma: aprognostic factor independent of Ras. Cancer Res. 1999;59:684–8.

16. Cheng YL, Lee SC, Harn H-J, et al. Prognostic prediction of theimmunohistochemical expression of p53 and p16 in resected non-small cell lung cancer. Eur J Cardiothorac Surg. 2003;23:221–8.

17. Esposito V, Baldi A, Vincenzi B, et al. Analysis of cell cycle regulatorproteins in non-small cell lung cancer. J Clin Pathol. 2004;57:58–63.

18. Hayashi H, Ogawa N, Ishiwa N, et al. High cyclin E and low p27/Kip1 expressions are potentially poor prognostic factors in lungadenocarcinoma patients. Lung Cancer. 2001;34:59–65.

19. Kawasaki M, Nakanishi Y, Kuwano K, et al. The utility of p53immunostaining of transbronchial biopsy specimens of lung can-cer: p53 overexpression predicts poor prognosis and chemoresistance

in advanced non-small cell lung cancer. Clin Cancer Res. 1997;3:1195–200.

20. Sanchez PA, Torres AJ, Iniesta P, et al. Prognostic significance of p53gene mutations in squamous cell carcinoma of the lung. Oncol R.1998;5:1129–33.

21. Tong J, Sun X, Cheng H, et al. Expression of p16 in non-smallcell lung cancer and its prognostic significance: a meta-analysisof published literatures Review Article. Lung Cancer. 2011;74:155–63.

22. Gorgoulis VG, Zacharatos P, Kotsinas A, et al. Alterations of the p16-pRb pathway and the chromosome locus 9p21–22 in non-small-celllung carcinomas: relationship with p53 and MDM2 protein expres-sion. Am J Pathol. 1998;153:1749–65.

23. Baldi A, De Luca A, Esposito V. et al. Tumor suppressors and cell-cycle proteins in lung cancer. Review article. Pathology ResearchInternational 2011;1-12

24. Esposito V, Baldi A, DeLuca A, et al. Prognostic role of the cyclin-dependent kinase Inhibitor p27 in non-small cell lung cancer. CancerRes. 1997;57:3381–5.

25. Kaye FJ. RB and cyclin dependent kinase pathways: defining adistinction between RB and p16 loss in lung cancer. Oncogene.2002;21:6908–14.

26. Lloyd RV, Erickson LA, Jin L, et al. p27kip1: a multifunctionalcyclin-dependent kinase inhibitor with prognostic significance inhuman cancers. Am J Pathol. 1999;154:313–23.

27. Schneider PM, Praeuer HW, Stoeltzing O, et al. Multiple molecularmarker testing (p53, C-Ki-ras, c-erbB-2) improves estimation ofprognosis in potentially curative resected non-small cell lung cancer.Br J Cancer. 2000;83:473–9.

28. Grob TJ, Kannengiesser I, Tsourlakis MC, et al. Heterogeneity ofERBB2 amplification in adenocarcinoma, squamous cell carcinomaand large cell undifferentiated carcinoma of the lung. Mod Pathol.2012;25(12):1566–73.

29. Cantero R, Torres AJ, Maestro ML, et al. Pronostic value of thequantified expression of p185 in non-small cell lung cancer. JThorac Cardiovasc Surg. 2000;119:1119–25.

30. Díez M, Pollan M, Maestro M, et al. Prediction of recurrence byquantification of p185 protein in non-small cell lung cancer tissue. BrJ Cancer. 1997;75:684–9.

31. Shi D, He G, Cao S, et al. Overexpression of the c-erbB-2/neu–encoded p185 protein in primary lung cancer. Mol Carcinog.1992;5:213–8.

32. Micke P, Basrai M, Faldum A, et al. Characterization of c-kit expres-sion in small cell lung cancer: prognostic and therapeutic implica-tions. Clin Cancer Res. 2003;9:188–94.

33. Meert AP, Martin B, Delmotte P, et al. The role of EGF-R expressionon patient survival in lung cancer: a systematic review with meta-analysis. Eur Respir J. 2002;20:975–81.

34. Gosney JR, Field JK, Gosney MA, et al. c-myc oncoprotein inbronchial carcinoma: expression in all major morphological types.Anticancer Res. 1990;10:623–8.

35. Barr LF, Campbell SE, Diette GB. c-Myc Suppresses the tumorige-nicity of lung cancer cells and down-regulates vascular endothelialgrowth factor expression. Cancer Res. 2000;60:143–9.

36. Liu D, Huang C, Kameyama K, et al. E-cadherin expression associ-ated with differentiation and prognosis in patients with non-small celllung cancer. Ann Thorac Surg. 2001;71:949–55.

37. Böhm J, Niskanen L, Kiraly K, et al. Expression and prognostic valueof α-, β-, and γ-catenins in differentiated thyroid carcinoma. J ClinEndocrinol Metab. 2000;85:4806–11.

38. Choi YS, Shim YM, Kim SH, et al. Prognostic significance of E-cadherin and ß-catenin in resected stage I non-small cell lung cancer.Eur J Cardiothorac Surg. 2003;24:441–9.

39. Kase S, Sugio K, Yamazaki K, et al. Expression of E-cadherin and b-catenin in human non-small cell lung cancer and the clinical signif-icance. Clin Cancer Res. 2000;6:4789–96.

Tumor Biol.

40. Kimura K, Endo Y, Yonemyra Y, et al. Clinical significance ofS100A4 and E-cadherin-related adhesion molecules in non-small celllung cancer. Int J Oncol. 2000;16:1125–31.

41. Nozawa N, Hashimoto S, Nakashima Y, et al. Immunohistochemicalα- and β-catenin and E-cadherin expression and their clinicopatho-logical significance in human lung adenocarcinoma. Pathol-ResPract. 2006;202:639–50.

42. Bröker LE, Huisman C, Span SW, et al. Cathepsin Bmediates caspase-independent cell death induced by microtubule stabilizing agents innon-small cell lung cancer cells. Cancer Res. 2004;64:27–30.

43. Esposito V, Baldi A, De Luca A, et al. Cell cycle related proteins asprognostic parameters in radically resected non-small cell lung can-cer. J Clin Pathol. 2005;58:734–9.

44. Hommura F, Dosaka-Akita H, Kinoshita I, et al. Predictive value ofexpression of p16INK4A, retinoblastoma and p53 proteins for the prog-nosis of non-small-cell lung cancers. Br J Cancer. 1999;81:696–701.

45. Takada M, Tada M, Tamoto E, et al. Prediction of lymph nodemetastasis by analysis of gene expression profiles in non-small celllung cancer. J Surg Res. 2004;122:61–9.

46. Donnem T, Lonvik K, Eklo K, et al. Independent and tissue-specificprognostic impact of miR-126 in nonsmall cell lung cancer:coexpression with vascular endothelial growth factor-A predicts poorsurvival.n2011;117:3193-3200

47. Zhu CQ, Shih W, Ling C-H, et al. Immunohistochemical markers ofprognosis in non-small cell lung cancer: a review and proposal for amultiphase approach to marker evaluation. J Clin Pathol. 2006;59:790–800.

48. Bremnes RM, Veve R, Gabrielson E, et al. High-throughput tissuemicroarray analysis used to evaluate biology and prognostic signifi-cance of the E-cadherin pathway in non-small-cell lung cancer. JCO.2002;20:2417–28.

49. Sulzer MA, Leers MPG, van Noord JA, et al. Reduced E-cadherinexpression is associated with increased lymph node metastasis andunfavorable prognosis in nonsmall cell lung cancer. Am J Respir CritCare Med. 1998;157:1319–23.

50. Ucvet A, Kul C, Gursoy S, et al. Valor pronóstico del receptor delfactor de crecimiento epitelial, factor de crecimiento endotelial vas-cular, E-cadherina, y p120 catenina en el carcinoma de pulmón nomicrocítico resecado. Arch Bronconeumol. 2011;47:397–402.

51. Nakashima T, Huang C, Liu D, et al. Neural-cadherin expressionassociated with angiogenesis in non-small-cell lung cancer patients.Br J Cancer. 2003;88:1727–33.

52. Mohamed S, Yasufuku K, Hiroshima K, et al. Prognostic implica-tions of cell cycle-related proteins in primary resectable pathologicN2 non-small cell lung cancer. Cancer. 2007;109:2506–14.

53. Sterlacci W, Tzankov A, Veits L, et al. A comprehensive analysis ofp16 expression, gene status, and promoter hypermethylation in sur-gically resected non-small cell lung carcinomas. J Thorac Oncol.2011;6:1649–57.

54. Sion-Vardy N, Freedman J, Lazarov I, et al. p27kip1 expression innon-small cell lung cancer is not an independent prognostic factor.Anticancer Res. 2010;30:3699–704.

55. Cheng YL, Lee SC, Harn HJ, et al. Prognostic prediction of theimmunohistochemical expression of p53 and p16 in resected non-small cell lung cancer. Eur J Cardiothorac Surg. 2003;23:221–8.

56. Esposito V, Deluca A, Baldi A, et al. Altered expression of p53 andRb tumor suppressor genes in lung cancer: relationship with survival.Int J Oncol. 1996;9:439–43.

57. Au NH, Cheang M, Huntsman DG, et al. Evaluation of immunohis-tochemical markers in non-small cell lung cancer by unsupervisedhierarchical clustering analysis: a tissue microarray study of 284cases and 18 markers. J Pathol. 2004;204:101–9.

Tumor Biol.