Aims To systematically compare different approaches for evaluating mucosal proliferative activity regarding their diagnostic role for delineating flat lesions of the bladder mucosa.

Methods 32 carcinoma in situ (CIS) and 31 flat non-CIS conditions (low-grade dysplasia and reactive atypia) of the bladder mucosa were assessed by two independent pathologists in two rounds in terms of their proliferative activity assessed by the mitotic counts on H&E-stained sections (mitoses per mm²) and immunohistochemically using the MIB-1 antibody and the mitosis marker phosphohistone H3 (PHH3). Two different approaches for immunoscoring (percentage of stained nuclei vs dichotomised height of mucosal staining considering lower half vs full-thickness marker expression) were applied. κ statistics were used to evaluate interobserver and intraobserver reproducibility.

Results Scoring the percentage of Ki67 expressing cell nuclei seems to be superior to dichotomisation of the height of mucosal staining as well as to PHH3 immunostaining and conventional mitotic counts in terms of delineating CIS from flat non-CIS conditions. This approach shows substantial (κ=0.62–0.65; p<0.001) interobserver and substantial to almost perfect (κ=0.67–0.83; p<0.001) intraobserver reproducibility.

Conclusions The MIB-1 antibody is a useful adjunct in the differential diagnosis of conventionally challenging flat lesions of the bladder mucosa. In particular, 16% or more Ki67 positive cell nuclei favours CIS over flat non-CIS conditions, whereas 15% or less Ki67 positive cell nuclei is supportive of non-CIS conditions. However, due to some important limitations of MIB-1 staining, the MIB-1 antibody should be used as a component of a panel.