THE CHARACTERIZATION OF BIOLOGICAL FEATURES IN DEDIFFERENTIATED LIPOSARCOMAS BY MEANS OF PRINCIPAL COMPONENTS AND DISCRIMINANT ANALYSES OF 25 COMPUTER-GENERATED VARIABLES FROM FEULGEN-STAINED NUCLEI AND HISTOLOGICAL SLIDES

Several groups of lipomatous tumors are not yet clearly characterized on the biological level. In order to attempt to classify the dediffere ntiated liposarcomas with respect to other types of malignant liposarc omas, 80 adipose tumors were submitted to the combination of two compu ter-assisted methodologies. These two methodologies consisted of i) th e determination of 25 variables, and ii) the analysis of the diagnosti c information contributed by these 25 variables by means of two comple mentary techniques, i.e. principal components and discriminant analyse s. The 25 variables were computed by means of image cytometry on Feulg en-stained nuclei and histological slides, quantitatively describing d istinct biological characteristics relating to morphonuclear (chromati n pattern) features (14 variables), nuclear DNA content distribution ( 9 variables), and tissue architecture pattern (2 variables). The 80 ad ipose tumors included 21 typical lipomas, 7 atypical lipomas (defined as extremity adipose tumors with a histopathological pattern of well-d ifferentiated liposarcomas), 16 retroperitoneal and 5 non-retroperiton eal abdominal well-differentiated liposarcomas, 9 dedifferentiated lip osarcomas, 8 myxoid (intermediate-grade tumor) and 14 pleomorphic (hig h-grade tumor) liposarcomas. The data strongly suggest that the dediff erentiated liposarcomas exhibit biological characteristics which are d istinct from those of low- and high-grade liposarcomas, but similar to those of intermediate ones. The results also show that typical and at ypical lipomas are two distinct biological entities. In contrast, the atypical lipomas and the well-differentiated retroperitoneal and non-r etroperitoneal liposarcomas exhibited a high number of similar biologi cal characteristics. Computer-assisted methods contribute valuable inf ormation to characterize lipomatous tumor biology.