Various cytometric techniques have been found useful for discriminating between benign and malignant conditions and between different degrees of malignancy. Cytometry has been applied to certain oral epithelial lesions in recent studies, but all of these studies focused on measuring the amount of DNA, and some also measured the nuclear area. For oral epithelial lesions, we propose image analysis as an appropriate cytometric method to analyze and measure variables related to nuclear shape, amount of DNA and chromatin distribution pattern. The image analyzing system and the nuclear variables to be assessed, which have been used successfully for analysis of breast and uterine tissue, will be applied to the analysis of red lesions for the first time. In particular, this study will attempt to answer the following questions: (1) Can image analysis be applied successfully to discriminate between oral white lesions with hyperplasia and white or white-and-red lesions with moderate or severe dysplasia? (2) If so, can we use the variables that discriminate most powerfully between these lesions to predict the malignant transformations or oral lesions? To answer these questions we conducted two pilot studies using the Letiz TAS Plus image analyzer. In study 1, we developed a model of three variables that could discriminate between oral epithelial hyperplasia and dysplasia with high accuracy. In study 2, we predicted the prognosis of 10 oral lesions showing dysplasia with 100% accuracy. Studies of the epidemiology of leukoplakia and its relation to oral cancer have pointed out that up to 17% of white lesions are dysplastic and that up to 15% of cases of oral squamous cell carcinoma are directly associated with oral leukoplakia. Analyzing and assessing selected variables of nuclei from these lesions using the image analyzer and appropriate statistical analyses should provide a model to discriminate between these lesions and to predict their malignant potential.
| Abdel-Salam, M; Mayall, B H; Chew, K et al. (1990) Which oral white lesions will become malignant? An image cytometric study. Oral Surg Oral Med Oral Pathol 69:345-50 |
