Optimal combination of immunohistochemical markers for subclassification of non-small cell lung carcinomas: A tissue microarray study of poorly differentiated areas

Background: Recently, studies have been reported on the optimal immunohistochemical markers for subclassification of non-small cell lung carcinoma (NSCLC). The main pitfall in subclassification of NSCLC is small specimen with poorly differentiated area. In this study, we added newly proposed markers (e.g. napsin A, SOX2) to conventional markers (p63, TTF-1, CK5/6, and CK7) and evaluated optimal combination for subtyping of NSCLC, primarily focusing on the poorly differentiated area. Methods: Eighty two resected NSCLCs with poorly differentiated areas were classified based on histologic findings. After histologic review, only poorly differentiated areas were selected and tissue microarrays were constructed to simulate small biopsy conditions. A total of 36 adenocarcinomas (ADCs), 38 squamous cell carcinomas (SQCCs), and 8 large cell carcinomas were included. All specimens were stained with TTF-1, napsin A, CK7, p63, CK5/6, and SOX2. Results: With respect to ADC, TTF-1 was positive in 19 of 36 cases (53%) and napsin A was in 25 of 36 (69%). Both markers were specific for ADC (100% and 98%, respectively). With TTF-1 and napsin A in combination, sensitivity increased to 75%. CK7 was sensitive (92%) but not specific marker (76%) for ADC. With respect to SQCC, p63 was positive in 35 of 38 cases (92%) and CK5/6 was in 23 (61%). Both markers were specific for SQCC (both 93%). With p63 and CK5/6 in combination, sensitivity increased slightly to 95%, but specificity was lower at 91%. SOX2 was specific (100%) but not sensitive marker (53%) for SQCC. Combinations did not substantially increase the diagnostic performance. Conclusion: A simple panel of napsin A. TTF-1, and p63 can be sufficient to reliably subclassify poorly differentiated areas of NSCLC. (C) 2011 Elsevier Ireland Ltd. All rights reserved.