Histologic tumour image synthesis based on a cellular automaton.

Cellular automata are suitable models for simulating tumour growth and invasion at the microscopic level. The results of the stimulation, however, is usually represented by a pattern of grid points, which is difficult to relate to real histological patterns by visual inspection. In this paper a 2-dimensional image synthesis module is presented, which provides a more realistic visualisation by texture mapping to tumour cell nuclei, cytoplasm and stromal elements, simulating a melanocytic tumour invading the reticular dermis. The enhancement by image synthesis facilitates a visual comparison of simulated images of real histological specimens.