The contribution of image cytometry and artificial intelligence-related methods of numerical data analysis for adipose tumor histopathologic classification

摘要: Thirty-five lipomatous tumors were quantitatively described using 47 variables generated by means of computer-assisted microscope analysis. Of these quantitative variables, 27 computed on Feulgen-stained specimens (25 cytologic and 2 histologic samples) and, the remaining 20, 8 related to vimentin S-100 protein immunostaining patterns other 12 glycohistochemical staining peanut agglutinin, succinylated wheat germ concavalin A agglutinin. The 35 included 6 atypical lipomas well differentiated, 5 dedifferentiated, myxoid, 10 pleomorphic liposarcomas. actual diagnostic value contributed each with respect tumor groups was determined decision tree technique, an artificial intelligence-related algorithm that forms part supervised learning algorithms. technique retained 8: i.e., tissue architecture-, DNA ploidy level-, morphonuclear-, 1 lectin histochemical-, immunostain-related variables. made use set up logical rules make it possible identify from differentiated liposarcomas, one hand, dedifferentiated liposarcomas those are pleomorphic, other. Thus, combination intelligence analyzing analysis samples can be considered expert system contributing significant information conventional diagnosis.