The challenge of classifying poorly differentiated tumours in the lung

Abstract

The lung is a common site for primary carcinomas as well as a wide range of metastatic carcinomas and other rarer primary and secondary tumours that clinically and radiologically may mimic lung cancer. The World Health Organization classification of lung tumours is based on the morphological appearances in resected specimens. This provides a widely applicable system for classification, which, although complex, can be applied universally. The vast majority of patients with lung cancer, however, do not have their tumours resected, and the application of the same classification system to small biopsy specimens is problematic due in large part to sampling. Recognition of this has led to the widespread use of the term 'non-small cell carcinoma'. Future developments in our understanding of lung cancer and the development of novel targeted therapies may mean that in the future classification will become based more on the molecular features of tumours that predict response to therapies and prognosis rather than simple morphology. The predictive validity of such an approach using small biopsy or cytology specimens, however, still requires to be established. The problem of tumour heterogeneity, which is problematic for morphological classification, may also pose similar challenges for molecular approaches.