Expression of proliferating cell nuclear antigen (PCNA) and Ki-67 in dysplasia in inflammatory bowel disease

Previous studies have revealed large variations in the interobserver agreement of dysplasia grading in inflammatory bowel disease. Therefore, we investigated the diagnostic value of two novel monoclonal antibodies (MIB 1 against Ki-67 and PC 10 against PCNA) in the detection of dysplasia.

A total of 62 biopsies were investigated and histologically classified as follows: 13 probably positive for dysplasia; 15 low-grade dysplasia; five high-grade dysplasia; and 15 inflammation without dysplasia and 14 normal controls. The percentage of positive Ki-67- or PCNA-stained nuclei (= labelling index) was determined in relation to the distribution throughout the mucosa.

In all biopsies PCNA-labelling index exceeded that of Ki-67-labelling index. In the superficial half of the crypt PCNA- and Ki-67-labelling indices in the biopsies with 'indefinite for dysplasia, probably positive' or low-grade dysplasia exceeded that of normal tissue (P < 0.001). However, an unequivocal discrimination between biopsies with 'indefinite for dysplasia, probably positive' or low-grade dysplasia and inflammation was not possible. PCNA- and Ki-67-labelling indices were significantly higher in high-grade than in low-grade dysplasia (PCNA 81.4% vs. 44.3%, Ki-67 54.8% vs 30.9%, P < 0.001). Most interestingly, labelling indices of both markers were significantly (P < 0.0001) higher in biopsies with high-grade dysplasia than with active inflammation in the superficial half of the crypt.

PCNA and Ki-67 are useful adjuncts in the diagnosis of high-grade dysplasia, because high-grade dysplasia can easily be distinguished from low-grade dysplasia or active inflammation if the distribution of the positive-stained cells within the mucosa is taken into account. Lower unspecific binding and lower influence on proliferation activity by inflammation prompts us to prefer Ki-67 (MIB 1).