Improving the reproducibility of the Gleason scores in small foci of prostate cancer--suggestion of diagnostic criteria for glandular fusion.

High upgrading rates of Gleason score 6 to 7 carcinomas between biopsy and radical prostatectomy specimens may be produced by change of fused glands of pattern 3 to pattern 4. Therefore, inter-observer reproducibility of fused and non-fused glands in biopsy specimens was analysed. Images of H&E stained slides of glands of carcinomas with Gleason score 6 and 7 (3 + 4) with and without glandular fusions with different lens magnification were analysed by 4 specialized genitourinary pathologists and 3 non-specialized pathologists. The definition of glandular fusion was a complete lack of any stromal fibres between a minimum of two glands and only one line of nuclei within the area of fusion. Overall agreement and inter-observer reproducibility of fused versus non-fused glands of non- and uro-pathologically specialized pathologists were lower in lens magnification of 50× in contrast to 200×. The inter-observer reproducibility of fused glands by specialized observer was higher than that of non-specialized pathologists. The results support the importance of strict but practicable criteria for the diagnosis of fused tumor glands in order to decrease the interobserver variability of Gleason scores, particularly in non-specialised pathologists.