ABSTRACT
Classification of rat hepatocellular proliferative lesions can vary between pathologists as the many qualitative histologic criteria have not been satisfactorily evaluated and ranked for prognostic value. Computer-assisted morphometry offers an objective method to evaluate certain cellular features. The Solt-Farber resistant hepatocyte model was used in this study to produce populations of rats with a full range of hepatocellular proliferative lesions. Cellular features within the lesions were then measured morphometrically and the data were analyzed by animal age and by subjective lesion diagnosis. The nuclear/cytoplasmic ratio followed by the cell area and nuclear area were found to be the most important parameters for separating microscopic foci and islands of cellular alteration, an early hyperplastic lesion, from other hepatocellular proliferative lesions. The coefficient of variation, as a relative measure of heterogeneity, increased in a linear manner for cell, nuclear and nucleolar areas as the animals aged and was significantly higher for cell and nuclear area in hepatocellular carcinoma compared to other hepatocellular proliferative lesions. Hepatocyte nodules (representing primarily late hyperplastic lesions) and persistent hepatocyte nodules (lesions with similarities to hepatocellular adenoma) could not be satisfactorily separated within the limits of this study. As these borderline lesions show a continuum of cytologic change, other features, such as architectural change, are necessary to satisfactorily classify them on a subjective basis. An alternative approach is to use discriminant functions derived from morphometric studies.
