In the diagnosis of melanocytic skin tumors, the assessment of the overall architectural pattern (silhouette) is often essential. We have previously shown by a computer simulation model that tumor patterns are likely to depend on the relative degrees of proliferation and motility of the tumor cells. In this study, we examined the morphological pattern of 12 cases each of nevocellular nevi, primary melanoma, and metastatic melanoma by image analysis. The patterns of the individual melanocytic skin tumors were compared statistically with patterns generated by computer simulation, which facilitates estimates of biological properties of the tumor cells. Additionally, mitotic counts were made to measure tumor cell proliferation. A comparison of the three diagnostic groups revealed that the cells of nevocellular nevi show a low degree of motility, which, however, still exceeds the very low degree of proliferation. Thus, an ldquo;invasiverdquo; pattern with numerous small nests and single cells at the base of the lesion is common. In primary malignant melanoma, tumor cell motility and tumor cell proliferation are significantly increased in various proportions, thus leading to varying morphological patterns. In metastatic melanoma, a strikingly elevated degree of proliferation exceeds the only slightly elevated degree of motility and leads to sharply demarcated, ldquo;expansiverdquo; lesions. To check the validity of the technical procedure, estimates of proliferation based on pattern analysis alone were compared with the results obtained by mitotic counts. There was a significant correlation, indicating that the assumptions of the computer model and the image analysis procedure are in fact applicable to real-life melanocytic skin tumors. The study shows that architectural features routinely observed in melanocytic skin tumors are related at least in part to certain biological properties of the tumor cells.
展开▼
