The magiscan image analyser as a diagnostic aid in cytology

Many reports in recent years have described the application of image analysis to cervical smear prescreening. This paper describes a complementary approach to automation in diagnostic cytology. In this approach the image analyzer is programmed to automatically scan a slide, measure preselected morphologic features and store the results. Subsequently, the operator selects combinations of measurements for interactive analysis. Using this technique the diagnosis of subtle deviations in cell morphology may be based on a quantitative analysis of cell morphology. We describe the implementation of these techniques on the Magiscan Image Analyser and illustrate the process of analysis using three smears. These smears were routinely diagnosed as negative in one case and as cervical intraepithelial neoplasia (CIN grade II-III) in two cases.