It is a general opinion that tissue-specific stem cells are present in adult tissues but their specific properties remain elusive. They are rare in tissues and heterogeneous; in addition, their identification and the characterization of their progeny has encountered technical difficulties. In particular, the existence of pancreatic stem cells remains elusive because specific markers for their identification are not available. We established a method for the isolation of a population of stem/progenitor cells from the human exocrine pancreas, and propose it as a model for other human compact organs. We also used markers that identified and finally characterized these cells. Spheroids with self-replicative potential were obtained from all specimens. The isolated population contained a subset of CD34+ CD45- cells and was able to generate, in appropriate conditions, colonies that produce insulin. We obtained evidence that most freshly isolated spheroids, when co-cultured with the c-kit positive neuroblastoma cell line LAN 5, produced a c-kit positive progeny of cells larger in their cytoplasmic content than the original spheroid population, with elongated morphology resembling the neuronal phenotype. We identified a novel predominant functional type of stem/progenitor cell within the human exocrine pancreas, able to generate insulin-producing cells and potentially non-pancreatic cells.