Intra-observer variabilities of MIB-1-LI results.

<p>Shown are in each panel the MIB-1-LI results for each pathologist during step 2 of the study (left end of each line, one line representing the values obtained by one pathologist) and during step 3, 4 months later (right end of each line). The grey zone delineated the 8% and 15% cut-offs used for calculaing Kappa values. The three rows represent the results obtained for the three different carcinomas (cases 1–3). The left two columns of panels show the results obtained by reading MIB-1-LI in areas that the pathologists selected themselves (self-selected) and the panels in the two right columns the values assessed in pre-defined areas (pre-defined). In each pair of columns, values were assessed by eyeballing (left) and by counting (right). As expected, Kappa values for intraobserver variability were slightly better in pre-defined areas than in self-selected areas, and, in analogy to the interobserver variability studies, better in eyeballed than in counted MIB-1-LI read-outs.