Cellular sociology applied to neuroendocrine tumors of the lung: quantitative model of neoplastic architecture

This paper reports on cellular sociology, which consists of modeling tissular architecture based on graph theory. Voronoi's diagram was chosen to build the models. This diagram derives from a cell neighborhood concept and generates parameters which objectively represent tissue architecture. Minimal spanning tree (MST) is probably the more frequently used among graphs and successfully discriminates different grades of pathological process. However, Voronoi's diagram is more comprehensive and a more complete representation of architecture with the advantage of stability. The lung neuroendocrine tumor classification is far from being consensual, especially for lesions which don't fall in with typical carcinoid and small cell carcinoma groups. By comparing architectural models of 20 neuroendocrine tumors of the lung, this work supports the morphologic spectrum concept of these tumors and also supports the recently proposed concept of large-cell neuroendocrine tumors of the lung. Finally, architectural parameters separate small-cell-lung carcinomas from neuroendocrine non-small-cell lung carcinomas.