Introduction: The emergence of treatments for non-small cell lung carcinoma (NSCLC) with differential efficacy and toxicity between subtypes has highlighted the importance of specific pathologic NSCLC subtyping. Most NSCLCs are inoperable, and pathologic diagnosis is made only on small tissue samples that are prone to diagnostic inaccuracy. In a substantial proportion of cases, standard morphology cannot specifically subtype the tumor, necessitating a diagnosis of NSCLC-not otherwise specified (NOS). Histochemical staining for mucin and immunohistochemical (IHC) identification of NSCLC subtype-associated markers could help predict the final subtype of resected NSCLCs diagnosed as NSCLC-NOS on preoperative bronchial biopsy samples.
Methods: Paraffin sections of 44 bronchial biopsy samples diagnosed as NSCLC-NOS were stained for mucin (Alcian blue/periodic acid Schiff) and thyroid transcription factor 1 by IHC-(markers of adenocarcinoma), and for S100A7, cytokeratin 5/6, high molecular weight cytokeratins, and p63 proteins-markers of squamous cell carcinoma. A predictive staining panel was derived from statistical analysis after comparing staining profiles with the final postsurgical NSCLC subtype. This panel was prospectively applied to 82 small biopsy samples containing NSCLC.
Results: True NSCLC subtype of undifferentiated NSCLC samples was best predicted using Alcian blue/periodic acid Schiff plus p63 and thyroid transcription factor 1 IHC, allowing specific subtyping in 73% of NSCLC-NOS cases with 86% accuracy. When applied prospectively, this staining panel showed 100% concordance with specific NSCLC morphologic subtyping in small biopsies.
Conclusion: This approach can facilitate treatment selection by accurately predicting the subtype in undifferentiated NSCLC biopsies, reducing to 7% the proportion of cases without a definite or probable histologic subtype.