ALGORITHM ANALYSIS OF LECTIN GLYCOHISTOCHEMISTRY AND FEULGEN CYTOMETRY FOR A NEW CLASSIFICATION OF NASAL POLYPOSIS

The aim of this study is to present a new classification of nasal poly ps. This classification is based both on morphologic criteria relating to morphonuclear features from isolated Feulgen-stained nuclei and on glycohistochemical characteristics from histologic slides submitted t o three lectins (peanut, wheat germ, and gorse seed agglutinins) and o ne neoglycoconjugate glycohistochemical stain. While the morphonuclear features (including 30 variables) relate essentially to chromatin pat tern, the glycohistochemical stains (including 16 variables) are linke d to the presence of specific carbohydrate moieties in cell membranes and cytoplasm. Forty-nine nasal polyps, including single polyps, diffu se polyposis, cystic fibrosis-related polyposis, and aspirin idiosyncr acy-related polyposis associated with asthma, were thus characterized. All the variables were obtained quantitatively by means of computer-a ssisted microscopy. Two complementary methods of data classification w ere used to determine the actual diagnostic value contributed by each quantitative variable, namely, discriminant analysis, which forms part of multifactorial statistical analysis, and the decision tree techniq ue, which is an artificial intelligence-related algorithm. The data so obtained show that our morphologic classification of nasal polyps fit s in with the classification of nasal polyps defined on the basis of c linical criteria.