Computer Analysis of Lymphocyte Images

Abstract

Research on computer analysis of microscopic images of cells concentrates at this time on three major applications. These are the automated recognition of cells from the hematopoietic system, the analysis of epithelial cells and the study of lymphocyte populations. Of these, the first two are clearly related to the immediate needs of the clinical laboratory. White blood cell differential counts are carried out at a rate of more than 100 million per annum in the United States alone. Research and development here has progressed from the first feasibility studies accomplished in the 1960s by Preston (1961) and Preston (1962), Prewitt and Mendelsohn (1966), Ingram et al. (1968), Young (1969), and Bacus (1970) to the commercially available automated white blood cell differential counting devices. For the analysis of epithelial cells, the major effort has concentrated on the clinical cytology of the female reproductive tract (Wied et al., 1976). Here, research toward the improvement of the diagnostic characterization of the material and extensive efforts to automate the prescreening for cervical cancer are underway. Also directed toward early detection and diagnosis of malignant disease, and particularly toward the extraction of prognostic clues, are research projects on image analysis of urothelial cells (Koss et al., 1975, 1977a,b, 1978a; Bartels et al., 1977c) and cells from the respiratory tract (Reale et al., 1978; Wied et al., 1979). The analysis of digitized images of lymphocytes at this time does not have such immediate clinical applications.