A DNA Ploidy Study Using Image Analysis in melanocytic Skin Lesions / 대한피부과학회지

ABSTRACT

BACKGROUND: In the cutaneou; neoplasms, especially melanocytic skin lesions, the degree of the nuclear atypia such as the hyp rchromatism, pleomorphism, and giantism is a very impor tant factor when determining the differntial diagnosis between the malignant. and the benign turnors and the prognosis. OBJECTIVE: So we measured and cornpared the nuclear atypia of the malignant melanoma, and acquired and congenital melanocytic nevi(AMN and CMN) using the computerized image analysis (IA). METHODS: Five micron sections were cut from each paraffin block. The sections were Feulgen-stained and image analysis was perforrned with an AIC image analysis system. In each case, 50 lymphocytes nuclei fnim the sarne specimen slide were measured as an internal diploid control. Each 100 cells of the MM, CMN and AMN were measured and are referred to in this report as the upper part and the lower part. The morphological parameters such as mean nuclear area, coefficient of variation(CV) of the nuclear area, roundness of the nucleus and CV of the roundness were measured. For the nuclear content, mean ploidy, CV of the mean ploidy ind 2c Deviation Index(2cI3I) were calculated and then the DNA histograms were produced. RESULTS: Nuclear area, CV of the nuclear area, mean ploidy, CV of the mean ploidy and 2cDI were found to be significantly reat.er for rnalignant melanoma than for CMN and AMN. The nuclear area of AMN is significantly lar ger than that of CMN and there were no significant differences between the uper and the lower parts of three melanocytic lesions in any of the measured parameters. CONCLUSION: The results of o ir study suggest that the IA of the melanocytic cells might reflect the biologic behavior rnore sensitively than do clinical or histologic criteria, therefore important information for differentiating thi benign and the malignant melanocytic lesions may be obtained by the DNA profile and morphomi try on Feulgen stained tissue specirnens using IA.