The clinical management of patients with colorectal polyps is mainly based on the histology of the removed lesions. Therefore, the histological diagnosis has a very important role in deciding the treatment and the uniform interpretation of the different pathology reports is essential. In spite of these relevant implications, there are only very few studies assessing the interobserver variability in such diagnosis and the concordance of reports among different examiners is not considered satisfactory. OBJECTIVE: to assess interobserver variability in the pathology reports in the diagnosis of colorectal polyps. METHOD: at the Department of Pathology of HCPA [Hospital de Clínicas de Porto Alegre] 230 slides of colorectal polyps were examined by four independent pathologists "blindly", that is, the diagnosis given by their colleagues was not known. Colorectal lesions were classified according to the diagnosis as polyp or invasive cancer and to the polyp type (adenomatous or hyperplasic). The histological type of the adenomas (tubular, tubulovillous and villous) and the grade of dysplasia (high or low) were also assessed. RESULTS: mean Kappa of the type of lesion was 0.794, which is considered moderate. The mean Kappa of 0.852 for the type of polyp is considered excellent concordance. Regarding the histology of adenomas, the mean Kappa was 0.291, considered weak. The assessment of the degree of dysplasia showed a regular Kappa of 0.420. CONCLUSION: the concordance rate among the four pathologists was considered to be moderate to excellent for the type of lesion and of polyp but there was great variability in the assessment of adenomas both for the histological type and for dysplasia, showing a weak to regular Kappa.
colon; polyps; adenomas; dysplasia; variability; Kappa
