ORIGINAL ARTICLE

Refining the Diagnosis and EGFR Status of Non-small CellLung Carcinoma in Biopsy and Cytologic Material, Using aPanel of Mucin Staining, TTF-1, Cytokeratin 5/6, and P63,

and EGFR Mutation Analysis

Andrew G. Nicholson, DM, FRCPath,*† David Gonzalez, PhD,‡ Pallav Shah, MRCP,§Matthew J. Pynegar, BSc,* Manjiri Deshmukh, FRCPath,* Alexandra Rice, FRCPath,*

and Sanjay Popat, MBBS, MRCP, PhD†�

Introduction: The dichotomization of non-small cell carcinoma(NSCLC) subtype into squamous (SQCC) and adenocarcinoma(ADC) has become important in recent years and is increasinglyrequired with regard to management. The aim of this study was todetermine the utility of a panel of commercially available antibodiesin refining the diagnosis on small biopsies and also to determinewhether cytologic material is suitable for somatic EGFR genotypingin a prospectively analyzed series of patients undergoing investiga-tion for suspected lung cancer.Methods: Thirty-two consecutive cases of NSCLC were first testedusing a panel comprising cytokeratin 5/6, P63, thyroid transcriptionfactor-1, 34�E12, and a D-PAS stain for mucin, to determine theirvalue in refining diagnosis of NSCLC. After this test phase, twofurther pathologists independently reviewed the cases using a re-fined panel that excluded 34�E12 because of its low specificity forSQCC, and refinement of diagnosis and concordance were assessed.Ten cases of ADC, including eight derived from cytologic samples,were sent for EGFR mutation analysis.Results: There was refinement of diagnosis in 65% of cases ofNSCLC to either SQCC or ADC in the test phase. This included 10of 13 cases where cell pellets had been prepared from transbronchialneedle aspirates. Validation by two further pathologists with varyingexpertise in lung pathology confirmed increased refinement andconcordance of diagnosis. All samples were adequate for analysis,and they all showed a wild-type EGFR genotype.Conclusion: A panel comprising cytokeratin 5/6, P63, thyroidtranscription factor-1, and a D-PAS stain for mucin increases diag-nostic accuracy and agreement between pathologists when faced

with refining a diagnosis of NSCLC to SQCC or ADC. These smallsamples, even cell pellets derived from transbronchial needle aspi-rates, seem to be adequate for EGFR mutation analysis.

Key Words: Lung, Cancer, EGFR, Immunohistochemistry, Trans-bronchial needle aspirate, Cytology, Non-small cell carcinoma,Adenocarcinoma, Squamous cell carcinoma.

(J Thorac Oncol. 2010;5: 436–441)

Until recently, subclassification of non-small cell lungcancer (NSCLC) into different histologic types had little

clinical significance, because treatments did not differ be-tween the variants. However, several drugs have becomeavailable in recent years with indication contingent on precisedefinition of tumor subtype, in relation to both efficacy andsafety. Thus, the multitargeted antifolate pemetrexed is con-traindicated in squamous cell carcinomas (SQCC) with sev-eral trials demonstrating no significant efficacy in thisNSCLC subtype,1–3 but superior efficacy in the adenocarci-noma (ADC) and large cell subsets over standard platinum-doublet chemotherapy. Bevacizumab development was dis-continued in patients with squamous cell tumors due to anover-representation of fatal pulmonary hemorrhage in thissubgroup,4 and squamous histology is now a contraindicationfor this drug. The EGFR-directed tyrosine kinase inhibitorserlotinib and gefitinib have superior efficacy in patients har-boring somatic EGFR mutation, which is most frequentlyobserved in ADCs.5 Moreover, histology subtype may alsooffer prognostic information.5 In addition to the above need,to better identify histologic NSCLC subtypes, biopsies ofadequate size are also required for somatic genotyping. Re-cent trial data have demonstrated superiority of gefitinib overcytotoxic chemotherapy in patients with NSCLC harboringsomatic EGFR mutation; whereas in those with wild-typeEGFR, first-line gefitinib is markedly detrimental comparedwith chemotherapy. Thus, EGFR genotyping is now indicatedif considering therapy with first-line gefitinib.6

The NSCLC diagnostic paradigm is further com-pounded by the advent of transbronchial needle aspiration

*Department of Histopathology, Royal Brompton Hospital, London; †Na-tional Heart and Lung Institute, Imperial College London, United King-dom; ‡Department of Molecular Diagnostics, The Institute of CancerResearch, Surrey; §Department of Respiratory Medicine, Royal Bromp-ton Hospital; and �Department of Medical Oncology, Royal MarsdenHospital, London.

Disclosure: The authors declare no conflicts of interest.Address for Correspondence: Andrew G. Nicholson, DM, FRCPath, Depart-

ment of Histopathology, Royal Brompton Hospital, Sydney St, LondonSW3 6NP, United Kingdom. E-mail: a.nicholson@rbht.nhs.uk

Copyright © 2010 by the International Association for the Study of LungCancerISSN: 1556-0864/10/0504-0436

Journal of Thoracic Oncology • Volume 5, Number 4, April 2010436

(TBNA)/endobronchial ultrasound (EBUS) and endoscopicultrasound (EUS) guided aspiration, which mean that manypatients are being diagnosed and staged in a single procedurewithout recourse to any form of biopsy.7 This combination ofevents is therefore potentially problematic as the volume ofdiagnostic tissue is becoming ever smaller, with greaterrequirement for refinement of diagnosis.

A potential facilitator in subdividing NSCLCs in smallbiopsies is immunohistochemistry. Although most data areretrospective and based on resection or tissue microarrayspecimens and not small biopsies, antibodies to epitopes suchas P63,8–10 cytokeratin (CK) 5/6,9,11 and 34�E1212,13 areknown to stain selectively SQCCs and similarly thyroidtranscription factor-1 (TTF-1) for lung ADCs.9,10,14,15 Wetherefore aimed to determine the utility of a panel of com-mercially available antibodies to refine the diagnosis ofNSCLC by immunohistochemistry, and also to determinewhether cytologic material is suitable for somatic EGFRgenotyping in a prospectively analyzed series of patientsundergoing investigation for suspected lung cancer.

MATERIALS AND METHODSPatients referred to the Royal Brompton hospital for

suspected lung cancer, whose biopsy procedure was confir-mative, were included in the study. The methodology forbronchoscopy in patients undergoing TBNA has been de-scribed previously.7 Samples obtained from EBUS and EUSwere also collected into saline. A separate container was usedfor each lymph node station and processed as follows: aspi-rated material was placed in saline, one container per station.Blood was removed via density gradient centrifugation usingLymphoprep (Axis-Shield PoC AS), during which specimenswere spun at 2000 rpm for 20 minutes, followed by cytospinpreparation, and up to four slides were reviewed per station.In positive cases, the remaining material was then placed in10% formaldehyde and subsequently spun into a cell pelletfor routine histologic processing. As a test phase, all sampleswere first reviewed by one pathologist (A.G.N.) and werediagnosed as NSCLC-not otherwise specified (NSCLC-NOS), SQCC, ADC, or small cell lung carcinoma (SCLC).Diagnosis of SQCC and ADC was based on identifying thecriteria for diagnosis of these tumor types in the 2004 WorldHealth Organisation (WHO) classification, noting the caveatthat these criteria are based on resection specimens.16 Thesewere the presence of keratinisation and/or intercellular digi-tation for SQCC and the presence of lepidic/acinar/papillaryarchitectures for ADC.

Positive biopsies and positive cell pellets were thensubjected to an ancillary panel of immunohistochemistry forTTF-1 (Dako M3575, 1:100 dilution), p63 (Dako M7247,1:2000), CK 5/6 (Dako M7237, 1:200), 34�E12 (DakoM0630, 1:150), with additional staining for broad spectrumCKs using MNF116 (Dako M0821, 1:50), and neuroendo-crine differentiation using CD56 (Vector VP-C360, 1:100)when appropriate. Staining for neutral mucin using a dia-stase-resistant periodic acid Schiff (D-PAS) stain was alsoundertaken. Each specimen was then reviewed a second time,

along with the ancillary data, by the same pathologist(A.G.N.), and a final diagnosis was made.

Based on data from the earlier test panel by A.G.N.,two further pathologists (one thoracic specialist [A.R.] andone general pathologist [M.D.]) independently assessed thecases in similar fashion, and then reassigned their diagnosesusing a validation panel of TTF-1, P63, CK 5/6, and a D-PASstain for mucin.

Ten cases diagnosed as ADC, eight derived from cellpellets, were then sent for molecular analysis to look forevidence of EGFR mutations, where samples were available.Genomic DNA was extracted from formalin-fixed, paraffin-embedded tissue sections (five sections 10 �m width for eachsample) using the QIAamp DNA formalin-fixed, paraffin-embedded kit (Qiagen, UK). EGFR mutation analysis wascarried out using the EGFR29 Mutation Kit (DxS Ltd, Manches-ter, United Kingdom) according to manufacturer’s instructions.This method can detect the 29 most common EGFR mutationsin exons 18 to 21, and the kit can detect 1% of mutant DNA ina background of wild-type genomic DNA. Positive controlswere included in every run for the most common EGFR muta-tions and all of them gave positive results.

RESULTSBetween May 2008 and March 2009, 38 consecutive

biopsies were confirmed as primary lung cancer by onepathologist (A.G.N.) in patients attending the Royal Bromp-ton Hospital and Royal Marsden Hospital multidisciplinaryThoracic Oncology service. Six of these were diagnosed asSCLC, all confirmed by positive staining with neuroendo-crine markers. Of the remaining 32 NSCLC cases, 12 wereendobronchial biopsies, five were core biopsies (lung � 4,liver � 1), 13 were TBNAs (blind, EBUS, and EUS guided)with subsequent cell pellets, and two were pleural effusionswith subsequent cell pellets.

Seven of 32 cases were diagnosed as ADC, of which62.5% stained with TTF-1, with negative staining for CK5/6and only one case showing focal staining with P63, whereas50% stained with 34�E12. Eight of 32 cases were diagnosedas SQCC, all of these being negative TTF-1 and all positivefor CK5/6, P63, and 34�E12. One of these showed veryoccasional staining for mucin but was still classified as SQCCas the tumor cells fulfilled WHO morphologic criteria for adiagnosis of SQCC. The remaining 17 cases were diagnosedas NSCLC-NOS, with 10 of the 13 TBNA/cell pellet speci-mens being within this group.

Subsequent to the primary immunohistochemistrypanel, a diagnosis of NSCLC-NOS was refined to ADC in 9of 17 patients, tumor cells being either TTF-1 positive (9 of9) and/or D-PAS mucin positive (3 of 8). These cases wereeither wholly negative or only focally positive for CK5/6(2 of 10) and P63 (2 of 10) (Figure 1), although there wasa greater degree and extent of positivity for 34�E12 (9 of9). Three of these cases were noted to have marked nuclearpleomorphism. In two patients, the diagnosis of NSCLC-NOS was refined to SQCC, these both being positive forCK 5/6, P63, and 34�E12 and negative for both TTF-1 and

Journal of Thoracic Oncology • Volume 5, Number 4, April 2010 Refining the Diagnosis and EGFR Status of NSCLC

Copyright © 2010 by the International Association for the Study of Lung Cancer 437

mucin expression, although one case was only focallypositive for CK5/6.

In six cases, the final diagnosis remained as NSCLC-NOS, four of which were cell pellets derived from TBNAs.This was due to a lack of additional tumor cells within the cellpellet to facilitate immunohistochemistry in two cases, aninability to distinguish confidently between contaminatingnative benign respiratory epithelial cells and tumor cellsshowing lesser degrees of cellular atypia in two cases, and nostaining other than for 34�E12 in two cases.

Therefore, 34�E12 was dropped from the validationpanel, because the results proved noncontributory to diagno-sis because of frequent staining of both ADCs and SQCCs.

After this prospective test phase, a second retrospectiveindependent review by two further pathologists (A.R. andM.D.), blinded to the initial data, was undertaken on all 38samples. On routine staining, there was agreement betweenthese two pathologists in 31 cases, classified by both as SCLC(n � 6), ADC (n � 7), SQCC (n � 4), and NSCLC-NOS (n �14). Seven of the remaining cases had discordant classifications,with four being called SQCC by one pathologist and NSCLC-NOS by the other (SQCC/NSCLC), and three ADC by onepathologist and NSCLC-NOS by the other (ADC/NSCLC).

After review of data from the ancillary panel, all caseswith initially agreed diagnoses of ADC and SQCC remained

unchanged. Of the 14 cases classified initially as NSCLC-NOS/NSCLC-NOS, four were reclassified as ADC by bothpathologists (ADC/ADC) (Figure 1), three as SQCC by bothpathologists (SQCC/SQCC) (Figure 2) and four to ADC byone pathologist (NSCLC/ADC). Three cases remained asNSCLC-NOS/NSCLC-NOS, all three samples being cell pel-lets derived from TBNAs. In the seven initially discordantcases, all four SQCC/NSCLC were reclassified by both pa-thologists as SQCC, and two of the three NSCLC/ADC casesby both pathologists as ADC (Figure 3).

Assessment of all data from all three pathologists showedno cases where there was a discordant diagnosis of ADC andSQCC in the same patient after ancillary investigation.

Ten cases with a diagnosis of ADC were subsequentlysent for gene mutation analysis. In all of these specimens,sufficient DNA was present to be able to detect activatingEGFR mutations down to a sensitivity of at least 10%(ranging from 1 to 10%). Sensitivity was calculated based onthe quantity and quality of the DNA used for EGFR mutationanalysis. None showed EGFR mutations (Table 1).

DISCUSSIONMost patients with NSCLC present with advanced

disease and require systemic therapy. For these, the di-

FIGURE 1. (A) A case of non-small cell lung carcinoma on TBNA of lymph node shows (B) undifferentiated tumor cells in thecell pellet that stain for (C) TTF-1, with reclassification as adenocarcinoma.

FIGURE 2. A case of non-small cell lung carcinoma on endobroncial biopsy is reclassified as squamous cell carcinomaafter immunohistchemistry. (A) H&E, (B) P63 positive (left side of field), and (C ) CK5/6 positive (left side of field). Notethe focal positive native epithelial cells in (B) and (C ) (right side of field) that need careful interpretation to avoid falsepositive diagnoses.

Nicholson et al. Journal of Thoracic Oncology • Volume 5, Number 4, April 2010

Copyright © 2010 by the International Association for the Study of Lung Cancer438

chotomization of NSCLC subtype into squamous andnonsquamous has become important in recent years, withthe advent of pemetrexed and bevacizumab therapy. Thistrend is continuing with other agents being developed with

differential efficacy by NSCLC subtype, for example in-sulin-growth factor receptor type 1 inhibitors show greatpromise in squamous subtype tumors.17

This study shows that a small immunohistochemistrypanel of TTF-1, CK 5/6, and P63, together with a mucin stain,refined diagnosis of NSCLC-NOS to either ADC or SQCC in65% of cases in the test phase. This included 6 of 13 samplestaken via TBNA, with subsequent cell pellet formation from theremaining fluid. This ancillary panel also produced greateragreement and refinement of diagnosis on independent valida-tion by two further pathologists at ends of the spectrum ofexpertise in pulmonary pathology. Of note, immunohistochem-istry produced no cases with conflicting data where ADC andSQCC were diagnosed in the same patient.

Although several antibodies are described as beingmarkers of squamous differentiation (P63, CK5/6, and34�E12) and adenocarcinomatous differentiation (TTF-1),there is very little published literature in relation to its usagein small biopsies in the lung, even more so with cytologypreparations, specifically TBNAs. Studies have shown thatTTF-1 and P63 are useful in distinguishing between SCLC

FIGURE 3. Validation by two pathologists shows greater concordance of diagnoses for either SQCC or ADC from 11 to 24out of 32 cases, through usage of an ancillary panel comprising thyroid transcription factor-1 (TTF-1), cytokeratin 5/6 (CK5/6),P63, and a diastase-resistant periodic acid Schiff (D-PAS) stain for mucin. Yellow boxes are diagnoses on routine staining atfirst review. Green boxes are diagnoses after review of ancillary panel. Adenocarcinoma (ADC)/ADC, squamous cell carcinoma(SQCC)/SQCC, non-small cell carcinoma-not otherwise specified (NSCLC-NOS)/NSCLC-NOS denotes agreed diagnoses by vali-dating pathologists of ADC, SQCC and NSCLC-NOS, respectively. SQCC/NSCLC-NOS denotes classification as SQCC by onepathologist and NSCLC-NOS by the other. ADC/NSCLC-NOS denotes classification as ADC by one pathologist and NSCLC-NOS by the other.

TABLE 1. EGFR Mutation Analysis on Small Biopsies

ID Sample Type EGFR Status Sensitivity (up to), %

1 TBNA pellet Wild type 1

2 TBNA pellet Wild type 1–5

3 TBNA pellet Wild type 1

4 TBNA pellet Wild type 1–5

5 Endobronchial biopsies Wild type 1–5

6 TBNA pellet Wild type 5–10

7 Endobronchial biopsies Wild type 5

8 Pleural eff pellet Wild type 1

9 TBNA pellet Wild type 1

10 Pleural eff pellet Wild type 1–5

EGFR, epidermal growth factor receptor; TBNA, transbronchial needle aspirate;Eff, effusion.

Journal of Thoracic Oncology • Volume 5, Number 4, April 2010 Refining the Diagnosis and EGFR Status of NSCLC

Copyright © 2010 by the International Association for the Study of Lung Cancer 439

and poorly differentiated SQCCs,9,10 although other havereported P63 staining in one third of ADCs on a study usingtissue microarrays.18 Our data suggest that the both theseantibodies are consistently of value in distinguishing ADCfrom SQC. One recent study in relation to a trial of bevaci-zimab reported that that P63 and CK5/6 were useful indistinguishing SQCC from ADC in fine needle aspirate spec-imens,19 data to support our findings. Our findings alsohighlight the value of staining for TTF-1 and the technicallymore simple investigation of looking for mucin in identifyingADC. Evidence of mucin production has long been recog-nized as a good marker of adenocarcinomatous differentia-tion16,20 but has often been forgotten in recent years withadvances in immunohistochemistry. With positive staining inapproximately 60% in pulmonary ADCs, this should not beforgotten as a useful test.

The prime criticism of this study is that resections arenot available for confirmation of classification. However, allof these antibodies are well described as being either markersof squamous differentiation8,10,11 or adenocarcinomatous dif-ferentiation14,15 in resection specimens. Furthermore, onerecent study of resection specimens diagnosed as ADC, SQCor large cell carcinoma, for whom an initial biopsy wasdiagnosed as NSCLC only, showed high specificities andsensitivities for the four markers in our panel in relation tocorrect assignment of diagnosis on biopsy to SQC or ADC.21

It is therefore reasonable to conclude that our data on biopsiesand also on tumor cells from TBNAs are a true reflection oftumor type, given that the current pathologic classification forlung neoplasms solely relates to resections,16 whereas 85% ofpatients only have biopsy material. Indeed an InternationalAssociation for the Study of Lung Cancer/American ThoracicSociety/European Respiratory Society-sponsored multidisci-plinary workshop on ADCs is ongoing with one of its remitsbeing to address this issue. Our data also demonstrate thatstaining for high molecular weight CKs (34�E12) showed amuch greater frequency and extent of staining across thespectrum of NSCLCs with many of the ADC samples beingpositive, suggesting limited use for this antibody in thiscontext. Others have however suggested that this antibodymay be useful in this context.22

A second issue is tumor heterogeneity in that, as in thisstudy, SQCCs may contain very occasional mucin vacuolesand no antibody is 100% specific. In addition, more centralADCs, which encompass the nonterminal respiratory unittype ADCs,23 will be negative for TTF-1, however stillpositive with a mucin stain in some cases. On this basis,although there is similarity in staining profiles betweenCK5/6 and P63 and overlap between TTF-1 positivity and thepresence of mucin within tumor cells in ADCs, we currentlyfavor the usage of all these tests. Further studies may refineour practice in due course.

We extended analysis of patients with a final diagnosisof ADC to assess whether biopsy material, especially fromcytologic/TBNA samples, are sufficient for somatic genotyp-ing. All 10 specimens analyzed contained sufficient DNA ofadequate quality and concentration, but no EGFR mutationswere detected. With the incidence of EGFR mutations rang-

ing from approximately 10% in populations outside the farEast24,25 this is likely to simply reflect a true absence giventhe small numbers analyzed, because the genotyping meth-odology used has a dilutional sensitivity lower limit ofapproximately 1%. However, our numbers are small andcontinued investigation of larger cohorts is therefore requiredbefore such samples can be viewed as routinely adequate fortherapeutic decision making.

Finally, although immunohistochemistry undoubtedlyrefined a diagnosis of NSCLC, there remained cases, espe-cially those derived from TBNA, where the final diagnosisremained as NSCLC-NOS due primarily to a lack of immu-nostaining precluding more accurate diagnosis, due to eithersmall volumes of tumor or heavy contamination by nativeepithelial cells in TBNAs. In addition, there were three caseswith marked pleomorphism that highlight issues in relation toextrapolation of a classification system based on resections tosmall biopsies. One could argue that such cases should be leftas NSCLC on morphology due to the likelihood of thembeing pleomorphic carcinomas, whereas their immunohisto-chemical profiles led to all three being preferably classified asADC, perhaps a better refinement of diagnosis in terms ofpotential systemic therapy options. This highlights the impor-tance of future classifications having input from disciplinesother than pathologists so that lung cancer classificationremains relevant to all user groups, and future reviews of theWHO classification will need to address such issues as andwhen they arise.

ACKNOWLEDGMENTSSP is in receipt of a Clinical Senior Lectureship Award

from the Higher Education Funding Council for England. SPand DCG also acknowledge NHS funding to the Royal MarsdenHospital/Institute of Cancer Research NIHR Biomedical Re-search Centre.

REFERENCES1. Scagliotti GV, Parikh P, von Pawel J, et al. Phase III study comparing

cisplatin plus gemcitabine with cisplatin plus pemetrexed in chemother-apy-naive patients with advanced-stage non-small-cell lung cancer.J Clin Oncol 2008;26:3543–3551.

2. Belani CP, Brodowicz T, Ciuleanu T, et al. Maintenance pemetrexed(Pem) plus best supportive care (BSC) versus placebo (Plac) plus BSC:A randomized phase III study in advanced non-small cell lung cancer(NSCLC). J Clin Oncol 2009;27.

3. Scagliotti GV, Park K, Patil S, et al. Survival without toxicity for cisplatinplus pemetrexed versus cisplatin plus gemcitabine in chemonaive patientswith advanced non-small cell lung cancer: a risk-benefit analysis of a largephase III study. Eur J Cancer 2009;45:2234–2235.

4. Johnson DH, Fehrenbacher L, Novotny WF, et al. Randomized phase II trialcomparing bevacizumab plus carboplatin and paclitaxel with carboplatinand paclitaxel alone in previously untreated locally advanced or metastaticnon-small-cell lung cancer. J Clin Oncol 2004;22:2184–2191.

5. Hirsch FR, Spreafico A, Novello S, et al. The prognostic and predictiverole of histology in advanced non-small cell lung cancer: a literaturereview. J Thorac Oncol 2008;3:1468–1481.

6. Mok TS, Wu YL, Thongprasert S, et al. Gefitinib or carboplatin-paclitaxelin pulmonary adenocarcinoma. N Engl J Med 2009;361:947–957.

7. Shah PL, Singh S, Bower M, et al. The role of transbronchial fine needleaspiration in an integrated care pathway for the assessment of patientswith suspected lung cancer. J Thorac Oncol 2006;1:324–327.

8. Camilo R, Capelozzi VL, Siqueira SA, et al. Expression of p63, keratin

Nicholson et al. Journal of Thoracic Oncology • Volume 5, Number 4, April 2010

Copyright © 2010 by the International Association for the Study of Lung Cancer440

5/6, keratin 7, and surfactant-A in non-small cell lung carcinomas. HumPathol 2006;37:542–546.

9. Kargi A, Gurel D, Tuna B. The diagnostic value of TTF-1, CK 5/6, andp63 immunostaining in classification of lung carcinomas. Appl Immu-nohistochem Mol Morphol 2007;15:415–420.

10. Wu M, Wang B, Gil J, et al. p63 and TTF-1 immunostaining. A usefulmarker panel for distinguishing small cell carcinoma of lung from poorlydifferentiated squamous cell carcinoma of lung. Am J Clin Pathol2003;119:696–702.

11. Chu PG, Weiss LM. Expression of cytokeratin 5/6 in epithelialneoplasms: an immunohistochemical study of 509 cases. Mod Pathol2002;15:6 –10.

12. Sturm N, Lantuejoul S, Laverriere MH, et al. Thyroid transcriptionfactor 1 and cytokeratins 1, 5, 10, 14 (34betaE12) expression in basaloidand large-cell neuroendocrine carcinomas of the lung. Hum Pathol2001;32:918–925.

13. Rossi G, Marchioni A, Milani M, et al. TTF-1, cytokeratin 7, 34betaE12,and CD56/NCAM immunostaining in the subclassification of large cellcarcinomas of the lung. Am J Clin Pathol 2004;122:884–893.

14. Ordonez NG. Thyroid transcription factor-1 is a marker of lung andthyroid carcinomas. Adv Anat Pathol 2000;7:123–127.

15. Kaufmann O, Dietel M. Thyroid transcription factor-1 is the superiorimmunohistochemical marker for pulmonary adenocarcinomas and largecell carcinomas compared to surfactant proteins A and B. Histopathol-ogy 2000;36:8–16.

16. Travis WD, Brambilla E, Muller-Hermelink H-K, et al., (Eds.). WHOClassification: Pathology and Genetics. Tumours of the Lung, Pleura,Thymus, and Heart. Lyons: IARC Press; 2004.

17. Karp DD, Paz-Ares LG, Novello S, et al. Phase II study of theanti-insulin-like growth factor type 1 receptor antibody CP-751,871 in

combination with paclitaxel and carboplatin in previously untreated,locally advanced, or metastatic non-small-cell lung cancer. J Clin Oncol2009;27:2516–2522.

18. Au NH, Gown AM, Cheang M, et al. P63 expression in lung carcinoma:a tissue microarray study of 408 cases. Appl Immunohistochem MolMorphol 2004;12:240–247.

19. Khayyata S, Yun S, Pasha T, et al. Value of P63 and CK5/6 indistinguishing squamous cell carcinoma from adenocarcinoma inlung fine-needle aspiration specimens. Diagn Cytopathol 2009;37:178 –183.

20. Travis WD, Colby TV, Shimosato Y, et al., and collaborators from 14other countries. Histological Typing of Lung and Pleural Tumors.WHO/IASLC Classification of Lung and Pleural Tumors, 3rd Ed.Berlin: Springer; 1999.

21. Loo P-S, Thomas SC, Nicolson MC, et al. Immunohistochemical diag-nosis of ‘non-small cell carcinoma-not otherwise specifiable’ in bron-chial biopsy specimens. J Thorac Oncol 2009;4:S340.

22. Rossi G, Pelosi G, Graziano P, et al. Review article: a reevaluation of theclinical significance of histological subtyping of non-small-cell lungcarcinoma: diagnostic algorithms in the era of personalized treatments.Int J Surg Pathol 2009;17:206–218.

23. Yatabe Y, Mitsudomi T, Takahashi T. TTF-1 expression in pulmonaryadenocarcinomas. Am J Surg Pathol 2002;26:767–773.

24. Zhu CQ, da Cunha Santos G, Ding K, et al. Role of KRAS and EGFR asbiomarkers of response to erlotinib in National Cancer Institute of CanadaClinical Trials Group Study BR.21. J Clin Oncol 2008;26:4268–4275.

25. Hirsch FR, Varella-Garcia M, Cappuzzo F, et al. Combination of EGFRgene copy number and protein expression predicts outcome for advancednon-small-cell lung cancer patients treated with gefitinib. Ann Oncol2007;18:752–760.

Journal of Thoracic Oncology • Volume 5, Number 4, April 2010 Refining the Diagnosis and EGFR Status of NSCLC

Copyright © 2010 by the International Association for the Study of Lung Cancer 441