COMPUTER-SIMULATION ANALYSIS OF MORPHOLOGICAL PATTERNS IN HUMAN MELANOCYTIC SKIN TUMORS

Tumour cell proliferation and particularly tumour cell motility are considered to be essential pre-requisites for invasive tumour growth. Despite abundant in vitro data on tumour cell motility, the behaviour of tumour cells in complex human tumour tissues is yet unknown. In this study, estimates of proliferation and motility are statistically derived from morphological tumour patterns in human melanocytic skin tumours. Two-dimensional, discrete, random computer simulations of tumour growth were carried out in order to determine the influence of tumour cell proliferation and motility on morphological patterns. A set of binary morphological criteria turned out to facilitate a significant estimate of the relative probabilities of motility and proliferation (CART analysis). When the same morphological criteria were applied to H & E stained slides of 45 melanocytic skin tumours, benign common nevi showed a predominance of motility, whereas primary and metastatic malignant melanoma revealed a predominance of proliferation. The direct assessment of the number of proliferating cells by Ki-67 staining shows a steep increase from benign nevi to primary and metastatic melanoma. These data provide first evidence that in benign common nevi the overall motility exceeds the very low degree of proliferation, whereas in malignant melanocytic tumours proliferation considerably exceeds tumour cell motility.