differential diagnosis of urothelial carcinoma in situ from non-neoplastic urothelia: analysis of ck20, cd44, p53 and ki67

background: flat urothelial lesions comprise a spectrum of morphologic changes ranging from reactive atypia to carcinoma in situ (cis). urothelial dysplasia and cis are associated with the recurrence and progression of urothelial carcinoma. distinguishing cis and dysplasia from reactive atypia based on histolopathogical features alone is often difficult. using different immunohistochemical markers such as cytokeratin 20 (ck20), cd44, p53, and ki-67 is recommended for differential diagnosis. the aim of this study was to evaluate the immunohistochemical pattern of these antibodies to differentiate different flat urothelial lesions.   methods: in this cross- sectional study, three groups of bladder biopsy specimens were evaluated: 20 samples with reactive urothelial lesions, 20 histologically diagnosed as cis, and 20 morphologically normal samples. immunohistochemical staining of ck20, p53, cd44 and ki-67 markers was performed on paraffin-embedded blocks. the groups were compared using chi square test, and the diagnostic value of the markers were evaluated with sensitivity, specificity, positive and negative predictive values.    results: ck20 was full thickness positive in 15 (75%) cis samples and negative in all samples of the normal and reactive groups (p< 0.001); cd44 was positive in 2 (10%) cases of the cis group and in 17 (85%) of the reactive group; this marker was negative in all the normal samples (p< 0.001). p53 was positive in 12 (60%) samples of the cis group and negative in all samples of the normal and reactive groups (p< 0.001). ki67 was positive in 13 (65%) samples of the cis group and 1 (5%) sample of the reactive group. this marker was negative in all samples of the normal group (p< 0.001).   conclusion: the results of this study revealed that ck20, cd44, p53 and ki67 are useful in distinguishing cis from reactive and normal samples. however, they should be used in a panel including at least three markers. correlation with the morphologic features is necessary.