A critical evaluation of AgNOR counting in benign naevi and malignant melanoma.

There is considerable variation in the quoted mean numbers of AgNORS per nucleus for benign melanonaevi and malignant melanomas. This is partly attributable to different approaches to AgNOR counting. This study summarizes our experience in devising an optimal technique for counting AgNORs. We show that it is essential to count intra-nucleolar AgNORs in addition to those lying outside the nucleolus to obtain clear separation of naevi from melanoma. Although this seems an onerous task, we further demonstrate that a maximum of only 30 nuclei need to be counted to obtain a mean AgNOR count per nucleus which is representative of the whole lesion. This compares with the arbitrary figure of 100 nuclei chosen by most workers. Only by optimizing and standardizing all aspects of the AgNOR technique including fixation, staining, and counting will mean AgNOR counts per nucleus become a useful quick, reproducible method which can be applied to lesions which pose diagnostic problems such as borderline melanocytic lesions.