Pulmonary neuroendocrine carcinomas in situ: do they exist?

Although precursor lesions of pulmonary squamous carcinoma, adenocarcinoma and carcinoid tumour are well known and accepted entities, precursor lesions of pulmonary neuroendocrine carcinomas, i.e. small-cell lung carcinoma and large-cell neuroendocrine carcinoma, have never been identified. A small body of literature documenting intraepithelial involvement in cases of pulmonary neuroendocrine carcinoma begs the question of whether the intraepithelial lesion represents de-novo intraepithelial neoplasia or invasion by underlying tumour. In this article, the literature on cases of invasive neuroendocrine carcinomas with an intraepithelial component showing a neuroendocrine immunophenotype is thoroughly and critically reviewed. The authors of the publications of cases of coexisting invasive and intraepithelial neuroendocrine carcinoma generally favour invasion as the explanation for the intraepithelial component. However, in practice, it is difficult or impossible to determine the direction of migration of tumour cells when there is both an invasive component and an intraepithelial component. Pulmonary neuroendocrine carcinomas in situ have been produced in genetically engineered mouse models. On the basis of the illustrations in the literature of human cases and the production of pulmonary neuroendocrine carcinomas in situ in genetically engineered mice, it seems likely that these tumours also exist in humans. However, their existence and acceptance require definitive proof. Studies should be undertaken to identify bronchial and other lung biopsies with in-situ carcinoma showing a neuroendocrine immunophenotype and a subsequent negative resection specimen and no follow-up evidence of invasive carcinoma. Documenting such cases would provide definitive proof for the existence of pulmonary neuroendocrine carcinomas in situ in humans.