Among lung neuroendocrine tumours (Lung-NETs), typical carcinoid (TC) and atypical carcinoid (AC) are considered separate entities as opposed to large cell neuroendocrine carcinoma (LCNEC) and small cell lung carcinoma (SCLC). By means of two-way clustering analysis of previously reported next-generation sequencing data on 148 surgically resected Lung-NETs, six histology-independent clusters (C1 → C6) accounting for 68% of tumours were identified. Low-grade Lung-NETs were likely to evolve into high-grade tumours following two smoke-related paths. Tumour composition of the first path (C5 → C1 → C6) was coherent with the hypothesis of an evolution of TC to LCNEC, even with a conversion of SCLC-featuring tumours to LCNEC. The second path (C4 → C2-C3) had a tumour composition supporting the evolution of AC to SCLC-featuring tumours. The relevant Ki-67 labelling index varied accordingly, with median values being 5%, 9% and 50% in the cluster sequence C5 → C1 → C6, 12% in cluster C4 and 50-60% in cluster C2-C3. This proof-of-concept study suggests an innovative view on the progression of pre-existing TC or AC to high-grade NE carcinomas in most Lung-NET instances.
Keywords: Cluster analysis; Lung; Neuroendocrine; Transition; Tumours.