RESUMO
<p><b>OBJECTIVE</b>To investigate the reproducibility of Gleason scores for prostate cancer.</p><p><b>METHODS</b>Based on the revised Gleason Scoring System of the International Society of Urological Pathology ( ISUP) , we analyzed the reproducibility and difference of Gleason scores in 49 cases of prostate cancer using the methods of combination and grouping.</p><p><b>RESULTS</b>The total reproducibility of Gleason scores among 15 pathologists was good (κ = 0.642), 62.2% by the combination method, the highest in Gleason 5 + 5 (81.2%) and 5 +4 (73.3%), then in Gleason 4 + 4 (67.5%), 3 + 3 (64.0%), 4 +3 (61.3%), and 3 + 4 (44.0%), and the lowest in Gleason 4 + 5 (38.9%) and 3 + 5 (33.3%). The total reproducibility of Gleason scores by the grouping method was 71.4%, the highest in Gleason 9-10 (84.9%) , then in Gleason 7 (76.7%) and 6 (64.0%), and the lowest in Gleason 8 (60.7%).</p><p><b>CONCLUSION</b>The reproducibility of Gleason scores remains to be further improved in prostate cancer, mainly concerning the understanding of Gleason 3 and 4 carcinoma.</p>