OUP user menu

Usefulness of Immunohistochemical and Histochemical Studies in the Classification of Lung Adenocarcinoma and Squamous Cell Carcinoma in Cytologic Specimens

Rebecca Ocque MD, Naobumi Tochigi MD, N. Paul Ohori MD, Sanja Dacic MD, PhD
DOI: http://dx.doi.org/10.1309/AJCPFKOLGL6PMOF3 81-87 First published online: 1 July 2011


Histologic subtyping of non–small cell lung carcinoma (NSCLC) is important because the efficacy of new treatments depends on tumor histologic features. We assessed the diagnostic accuracy of classification of lung adenocarcinoma and squamous cell carcinoma (SCC) on cytologic and biopsy specimens based on cytomorphologic studies alone or in combination with ancillary studies compared with resection specimens. Compared with adenocarcinoma, the diagnosis of SCC was based more often on cytomorphologic studies alone (139/185 [75.1%] vs 107/263 [40.7%]). Significantly increased use of immunohistochemical studies in cytology was noted after introduction of targeted lung carcinoma therapies (22/156 [14.1%] for adenocarcinoma and 5/46 [11%] for SCC from 2000–2004 vs 134/156 [85.9%] for adenocarcinoma and 41/46 [89%] for SCC from 2005–2010). Use of immunohistochemical studies resulted in increased diagnostic accuracy for adenocarcinoma (56% [44/78] from 2000–2004 vs 83.2% [154/185] after 2005) but not for SCC (77% [57/74] before 2004 vs 73.9% [82/111] from 2005–2010). Adenocarcinoma showed high expression of cytokeratin (CK)7 (146/146 [100%]), thyroid transcription factor-1 (131/152 [86.2%]), surfactant A (29/36 [81%]), and periodic acid–Schiff with diastase (69/86 [80%]). All SCCs were positive for CK5/6 and p63. Use of immunohistochemical studies on cytologic cell blocks may improve classification of NSCLC.

Key Words:
  • Lung
  • Adenocarcinoma
  • Squamous cell carcinoma
  • Immunohistochemistry
  • Cytology

Pathologists and oncologists have historically focused on discrimination between small cell lung carcinoma (SCLC) and non-SCLC (NSCLC) that included a large spectrum of histologically different tumor types. This was certainly a simplified view of lung cancer and reflected the clinical practice of the time and the rather limited efficacy of chemotherapy regimens for treatment of NSCLC. Hence, distinction between different subtypes of NSCLC, particularly adenocarcinoma and squamous cell carcinoma (SCC), was considered irrelevant for the management and treatment of lung cancer. Lately, this approach has been challenged after the results of several clinical trials identified lung tumor histologic features as a potential prognostic factor and predictor of clinical outcome and response to chemotherapy.

Several studies reported that patients with advanced lung carcinomas treated with standard palliative care with a combination of a platinum-based agent or with combination cytotoxic chemotherapy regimens showed better clinical outcome if the tumor was an adenocarcinoma.15 Georgoulias et al6 reported that patients with adenocarcinoma had significantly greater response to gemcitabine plus docetaxel, whereas patients with nonadenocarcinoma exhibited significantly greater response to cisplatin plus docetaxel. Similarly, extended survival was reported in patients with tumors with nonsquamous histologic features treated with induction chemotherapy regimens before radiation.7,8 The most consistent evidence of an association between histologic features and response to therapy has been reported in patients with lung adenocarcinoma treated with EGFR (epidermal growth factor receptor) tyrosine kinase inhibitors.911 In contrast, treatment with VEGF (vascular endothelial growth factor)–targeted recombinant humanized monoclonal antibodies, such as bevacizumab, resulted in fatal hemoptysis in patients with SCCs.12,13

As a result of these observations, oncologists are actively asking pathologists to subclassify NSCLC, particularly to at least separate SCC from non-SCC whenever possible. This distinction could be very challenging because most lung carcinomas, particularly advanced cases, are diagnosed on small cytologic or biopsy specimens. Pathologists are usually able to correctly distinguish between SCLC and NSCLC and to precisely classify well-differentiated or moderately differentiated SCC or adenocarcinoma. However, about one third of the cases—usually advanced, nonresectable cancers—fall into the category of poorly differentiated NSCLC without morphologic evidence of squamous, glandular, or neuroendocrine differentiation. As a result, ancillary techniques such as immunohistochemical analysis have been frequently incorporated into the diagnostic workup of many cytologic and small biopsy specimens. Many studies have addressed sensitivity and specificity of immunohistochemical markers in the separation of the various subtypes of NSCLC.1419 The interpretation of immunohistochemical studies on limited diagnostic material may be challenging, and misinterpretations are possible.

The aim of this study was to assess the diagnostic accuracy of classification of lung adenocarcinoma and SCC on small cytologic and biopsy specimens in real time based on cytomorphologic studies alone or in combination with ancillary studies. We also compared the diagnostic performance in the periods before and after the implementation of targeted therapies for the treatment of lung cancer. The reported interpretation of immunohistochemical results and their impact on diagnostic accuracy also were evaluated.

Materials and Methods

A total of 448 diagnostic preoperative cytologic specimens, including transthoracic computed tomography–guided fine-needle aspiration (FNA) specimens, bronchial brushings, and bronchial washings of a primary lung adenocarcinoma (263 cases [58.7%]) and SCC (185 cases [41.3%]), from patients who underwent surgical resection of the tumors at the University of Pittsburgh Medical Center, Pittsburgh, PA, from January 2000 to January 2010 were included in the study. Large cell carcinomas, large cell neuroendocrine carcinomas, small cell carcinomas, carcinoids, and adenosquamous carcinomas were excluded from the study owing to limited material.

Histologic classification of the tumor on the cytologic and resection specimens reported in the pathology electronic records was documented. Similarly, reported results of histochemical studies (mucicarmine and periodic acid–Schiff [PAS] with diastase) and immunohistochemical studies (cytokeratin [CK]7, CK20, thyroid transcription factor [TTF]-1, surfactant protein A, CK5/6, and p63), when available, were recorded.

Histologic diagnoses based on cytologic specimens were compared with those of the corresponding surgical resection specimens (wedge resections, lobectomies, and pneumonectomies). Classification accuracy of cytologic specimens was defined as histologic concordance between the cytologic classification and the final surgical resection diagnosis. Misclassification was defined as histologic discordance between the cytologic classification and the final surgical resection diagnosis. CK5/6 and p63 were used as immunohistochemical markers for SCC, and TTF-1 and surfactant protein A were markers for adenocarcinoma.

Statistical Analysis

Comparisons of the proportions of variables within diagnostic categories were performed by using the Fisher exact test. Statistical significance was defined as a P value of .05 or less, and all tests were 2-tailed. Sensitivity, specificity, and positive predictive value were calculated.


Cytologic Findings

Of 263 adenocarcinomas, the cytologic diagnosis was made based on cytomorphologic studies alone in 107 cases (40.7%), whereas ancillary studies, including immunohistochemical and/or histochemical studies, were used in 156 cases (59.3%). Of 185 SCCs, the cytologic diagnosis was made based on cytomorphologic studies alone in 139 cases (75.1%). Ancillary studies were used less frequently in cases of SCC than in adenocarcinoma (46 SCC [24.9%]; P = .001). Although cytologic cell blocks were available in all cases included in the study, there was a significant difference in the frequency of ancillary studies on cytologic specimens in both diagnostic groups before and after the introduction of targeted therapies for lung cancer in clinical practice after 2004. Between 2000 and 2004, only 22 (14.1%) of 156 adenocarcinomas and 5 (11%) of 46 SCCs were subjected to ancillary studies. In contrast, ancillary diagnostic studies were used in 134 adenocarcinomas (85.9%) and 41 SCCs (89%) between 2005 and 2010.

Diagnostic Performance

Histologic correlation of lung adenocarcinoma and SCC diagnoses on cytologic and surgical resection specimens is summarized in Table 1. Correct classification was made in 177 cases of adenocarcinoma (67.3%) and 140 cases of SCC (75.7%). The overall classification accuracy for adenocarcinoma was 56% (44/78) between 2000 and 2004 and 83.2% (154/185) between 2005 and 2010 (P = .0001). In contrast, there was no significant difference in overall classification accuracy for SCC: 57 (77%) of 74 cases between 2000 and 2004 and 82 (73.9%) of 111 cases between 2005 and 2010 (P = .7290) Figure 1.

View this table:
Table 1

The percentages of cases classified correctly on the basis of cytomorphologic studies alone were 56.1% (60 cases) for adenocarcinoma and 76.3% (106 cases) for SCC. When immunohistochemical/histochemical studies were used, the classification accuracy increased to 75.0% (117 cases) for adenocarcinoma (P = .002) but remained essentially unchanged at 74% (34/46 cases) for SCC (P = .843).

The overall number of adenocarcinomas classified as NSCLC was 82 cases (31.2%) with no significant difference in frequency of an NSCLC diagnosis between the 2 periods. Between 2000 and 2004, 41% (32/78) of adenocarcinomas were classified as NSCLC on cytologic specimens, and 27% (50/185) were classified as NSCLC between 2005 and 2010 (P = .2448). Similarly, there was no significant difference in the frequency of SCCs classified as NSCLC (14/74 [19%] between 2000 and 2004 vs 17/111 [15%] between 2005 and 2010; P = .5508). The number of adenocarcinomas classified as NSCLC by cytomorphologic studies alone was 54% (44/82 cases). Use of immunohistochemical/histochemical studies significantly reduced the number of adenocarcinomas classified as NSCLC (36 cases [23.1%]; P = .0026). In contrast, the number of SCCs classified as NSCLC by cytomorphologic studies alone (25 cases [18.0%]) was not significantly different from the number of SCC cases in which immunohistochemical/histochemical studies were used (10/46 [22%]; P = .5051).

The number of misclassified adenocarcinoma cases was the same for cases classified by cytomorphologic studies alone and cases in which immunohistochemical/histochemical studies were used (3 cases in each group; P = .6898). All adenocarcinomas misclassified by cytomorphologic studies alone were diagnosed as SCC. For the adenocarcinoma cases in which immunohistochemical/histochemical studies were used, 2 cases were misclassified as SCC and 1 case was misclassified as small cell carcinoma. The misclassified SCCs were diagnosed as adenocarcinomas in both groups. Although the overall number and proportion of SCCs in the immunohistochemical/histochemical studies group was slightly lower than in the group diagnosed by cytomorphologic studies alone, that difference did not reach statistical significance (P = .7331).

Figure 1

Diagnostic accuracy for adenocarcinoma (ADC) and squamous cell carcinoma (SCC) before and after introduction of targeted therapies for lung cancer in clinical practice.

Ancillary Studies

Table 2 shows the results of immunohistochemical/histochemical studies performed on the cell-block specimens.

There was large variability in the number of cell-block specimens stained with different antibodies, which reflects difficulties in working with a very limited amount of diagnostic tissue. All adenocarcinomas showed positive staining for CK7, followed by 86.2% of cases positive for TTF-1 and 81% of cases (20/36) positive for surfactant A. Similarly, 80% of the cases were positive for PAS with diastase (69/86), but only 65% were positive for mucicarmine (56/86). Surprisingly, 68% of SCC cases were reported to be positive for CK7 (28/41) and 21% for TTF-1 (9/43). In contrast, no surfactant A positivity was reported in SCCs. All of the tested cases of SCC were positive for CK5/6 and p63. The adenocarcinomas showed 21% (5/24) and 31% (13/42) positivity for CK5/6 and p63, respectively. The individual sensitivity, specificity, and positive predictive value for each marker were calculated separately and are summarized in Table 3.

View this table:
Table 2


The classification of lung carcinoma on cytologic specimens is becoming one of the most challenging areas of diagnostic cytopathology. Unfortunately, 70% to 80% of patients with lung cancer will not undergo definitive surgery, and the final diagnosis depends on cytologic or biopsy specimens only. The development of targeted therapies, of which the efficacy and toxic effects are linked to the histologic type of lung carcinoma, has resulted in a different diagnostic approach to lung carcinoma on small specimens. Morphologic examination alone is sufficient in cases of well-differentiated or moderately differentiated carcinoma. However, the accurate classification of NSCLC is often challenging and often not reproducible with poorly differentiated tumors, cases with limited diagnostic material, and morphologically heterogeneous tumors such as adenocarcinoma. In such cases, immunohistochemical analysis may be a useful ancillary tool. While most reports advocate the use of antibody panels for differentiating adenocarcinoma from SCC, many laboratories are using different approaches depending on technical abilities and the experience of pathologists. The majority of the published studies on this subject were designed to assess the expression of immunohistochemical markers in morphologically diagnosable adenocarcinoma or SCC in resection or cytologic specimens.14,18,2027 However, interpretation challenges and reporting of the results of immunohistochemical assays performed on limited diagnostic samples in clinical practice were not addressed.

View this table:
Table 3

Objective cytomorphologic criteria for adenocarcinoma and SCC are well established, and our study shows that these criteria were used more often solely without immunohistochemical/histochemical studies before the introduction of targeted therapies in clinical practice. As indicated in our report, only about 14% of adenocarcinomas and 8% of SCCs were subjected to immunohistochemical studies as a part of the diagnostic workup in the early 2000s, whereas about 86% of adenocarcinomas and 89% of SCCs had immunohistochemical studies performed after 2004. An interesting finding was that immunohistochemical studies did not increase the diagnostic accuracy of SCC. In contrast, a significant improvement in diagnostic accuracy was made with adenocarcinoma. This observation reflects the cytomorphologic heterogeneity of lung adenocarcinoma and the lack of reproducibility among pathologists in the classification of lung adenocarcinoma, particularly for poorly differentiated and solid types based on morphologic studies alone. Furthermore, immunohistochemical studies helped classify a significant portion of NSCLCs as adenocarcinomas.

Recently, the use of the generic term NSCLC, not otherwise specified, has been discouraged. The results from our institution reported in this study demonstrate a trend toward less frequent use of this term in recent years, but the overall difference was not significant despite efforts to use ancillary studies. There are several possible explanations for this observation, but the 2 most important would probably include tumor sampling and differentiation. Coghlin et al28 recently reported that significantly fewer tumor cells were present in bronchial biopsy specimens when the diagnosis of NSCLC, not otherwise specified, was made in comparison with other more specific tumor diagnoses. The size and tumor cellularity of the sample were influenced by the tumor location (central vs peripheral) and the experience of the bronchoscopist. Neither of these 2 deficiencies can be compensated by ancillary studies or diagnostic skills of pathologists. On the other hand, poorly differentiated carcinomas can be difficult to classify even on large resection specimens. The heterogeneity of immunostaining of lung carcinomas on resection specimens, particularly adenocarcinoma, is well known. Pathologists may be faced with even greater interpretation uncertainties with small specimens. Many poorly differentiated tumors may not show specific differentiation or demonstrate specific tumor marker expression, and, therefore, it is likely that hypocellular specimens from these tumors would remain unclassified despite ancillary studies. The question remains whether immunohistochemical studies should be used in such cases or if it is better to save the tissue for predictive marker testing.

Recent studies have proposed that the combined use of antibodies for adenocarcinoma (TTF-1 and CK7) and SCC (p63 and CK5/6) provides very reliable distinction between these 2 NSCLC subtypes.29 However, our study shows that when very limited diagnostic material is available, the interpretation may be subjective and difficult. All of our adenocarcinomas were positive for CK7, which is consistent with the findings of other studies.29,30 However, in contrast with the findings of other studies, CK7 positivity was reported in a disproportionately large number of SCCs. This unexpected finding can be explained by the subjective nature of interpreting immunohistochemical stains. In contrast with research studies, eg, estrogen receptor and progesterone receptor staining for breast carcinomas, percentages and intensity of staining of various markers for lung carcinomas are rarely included in the clinical reports. Therefore, the cases with faint and focal staining may be dismissed as negative by some pathologists but interpreted as positive by others. Furthermore, pathologists may dismiss aberrant antigen expression and base the final diagnostic decision on more convincing staining results and morphologic impressions.

TTF-1 expression has been reported in 70% to 85% of resected lung adenocarcinomas and between 60% and 90% of morphologically diagnosable adenocarcinomas on small cytologic or biopsy specimens.23,24,31,32 Our results for TTF-1 and surfactant A expression are very similar to those in previous reports. Although TTF-1 positivity in SCC was reported by some authors, the overall impression is that it should be absent in pulmonary SCC.24,26,31 In our review, about 20% of SCCs were reported as positive, most commonly peripheral SCC diagnosed by FNA. We believe that this observation is related to specimen heterogeneity. Tissue samples obtained by FNA or bronchoscopic biopsy represent a collection of different tissue types that, in addition to tumor, often contain benign bronchial mucosa, inflammatory cells, fibrotic tissue, and reactive pneumocytes. Pneumocytes adjacent to or admixed with the tumor may show significant cytologic atypia, which may lead to erroneous interpretation of TTF-1 positivity in tumor cells. In contrast with the findings of Kaufmann and Dietel,23 surfactant A, when compared with TTF-1, showed better specificity and positive predictive value in our study; none of the SCC samples were positive. These results should be interpreted with caution because of the relatively small number of cases studied with the surfactant A antibody. Similar to report by Loo et al,16 our study suggests that PAS with diastase may be considered a cost-effective alternative to TTF-1 or surfactant A for the diagnosis of adenocarcinoma.

Similar to prior studies, p63 and CK5/6 were excellent markers for SCC in our study.14,15,19 Comparable to the findings of Au et al,20 who studied a tissue microarray constructed from 405 surgically resected NSCLCs, our study showed about 30% of adenocarcinomas to be positive for p63. It should be kept in mind that p63 expression in adenocarcinoma tends to be focal and weak. In contrast, SCC shows staining that is usually diffuse and strong and most of the time not influenced by tumor grade.33 Positivity for p63 also occurs in bronchial reserve cells, which should not be mistaken for tumor cells showing positivity.34 It is interesting that immunohistochemical/histochemical studies did not improve the diagnostic accuracy for SCC in our study. This finding is in contrast with the findings of Jorda et al35 who reported an increased yield of detection of SCC from 35% to 88% by using p63 immunohistochemical staining. The reason for such discrepancy is uncertain, but it could be influenced by antibody type, specimen type, and pathologist experience.

Our study demonstrated a significant increase in the use of immunohistochemical and histochemical assays in routine cytopathology practice, leading to improved classification of adenocarcinoma but not SCC. The interpretation of immunohistochemical studies could be challenging on cell blocks, and misinterpretations are possible. Although the use of NSCLC in the diagnosis of lung carcinoma has been strongly discouraged, we found a significant proportion of undifferentiated or poorly differentiated tumors that, despite the use of ancillary studies, will remain unclassifiable. Similarly, a new International Association for the Study of Lung Cancer recommendation for the classification of adenocarcinoma discourages the use of the term nonsquamous carcinoma in pathology reports.36

From the perspective of treatment of NSCLC, the distinction between adenocarcinoma and SCC is most important for our clinical colleagues. Hence, information influencing therapeutic decision making may be obtained by other methods. Further studies are needed to determine whether ancillary studies for precise classification are more or less important than other potential molecular studies for prognostication of therapeutic response. Furthermore, ancillary studies should not be used for additional diagnostic reassurance in cases in which morphologic studies are sufficient for NSCLC subtyping.


Upon completion of this activity you will be able to:

  • list a panel of antibodies most useful in classification of non–small cell carcinoma of the lung.

  • apply a limited, but specific, panel of antibodies in diagnostic workup of lung specimens with only small samples of neoplastic cells.

The ASCP is accredited by the Accreditation Council for Continuing Medical Education to provide continuing medical education for physicians. The ASCP designates this journal-based CME activity for a maximum of 1 AMA PRA Category 1 Credit ™ per article. Physicians should claim only the credit commensurate with the extent of their participation in the activity. This activity qualifies as an American Board of Pathology Maintenance of Certification Part II Self-Assessment Module.

The authors of this article and the planning committee members and staff have no relevant financial relationships with commercial interests to disclose.

Questions appear on p 154. Exam is located at www.ascp.org/ajcpcme.


  1. 1.
  2. 2.
  3. 3.
  4. 4.
  5. 5.
  6. 6.
  7. 7.
  8. 8.
  9. 9.
  10. 10.
  11. 11.
  12. 12.
  13. 13.
  14. 14.
  15. 15.
  16. 16.
  17. 17.
  18. 18.
  19. 19.
  20. 20.
  21. 21.
  22. 22.
  23. 23.
  24. 24.
  25. 25.
  26. 26.
  27. 27.
  28. 28.
  29. 29.
  30. 30.
  31. 31.
  32. 32.
  33. 33.
  34. 34.
  35. 35.
  36. 36.
View Abstract