Immunocytochemical Staining for Cytokeratin 20 in Urine Cytology To Detect Neoplasia: Comparison with FISH Analysis.
Katherine A Devitt, Gladwyn Leiman, Jeannette Mitchell. Univ of Vermont College of Medicine, Burlington
Background: Urine cytopathology, widely used in screening for urothelial neoplasia, has high specificity (95%) but low sensitivity (40-80%). At this institution, cases are sent for fluorescent in situ hybridization (FISH) analysis by physician request, predominantly on cases classified as atypical or suspicious on cytologic evaluation; FISH is expensive and time-consuming. Recent literature suggests that cytokeratin 20 (CK20) immunohistochemical staining can identify urothelial neoplasia. CK20 staining is less expensive and more widely available than FISH. CK20 immunocytochemical staining and FISH results in 100 consecutive exfoliative cytopathology cases are correlated in this study.
Design: One hundred consecutive, archived Papanicolaou-stained urine cytology cases previously sent for FISH analysis were retrieved. Atypical cell groups were photographed prior to restaining with CK20 (Ks20.8, LabVision). Blinded to FISH results, slides were examined for immunostaining pattern and categorized as positive, negative, or indeterminate. CK20 and FISH results were correlated using simple statistics.
Results: Cytologic results on cases sent for FISH comprised 70 atypical, 18 negative, 7 suspicious and 5 malignant cases. FISH testing was negative in 71, positive in 20 and suspicious in 8 cases. One case had insufficient material for FISH. CK20 staining was negative in 72 cases, positive in 26. Two cases with indeterminate results were not included in analysis. The results of FISH analysis were used primarily as the gold standard for comparison. In discordant cases and those with suspicious FISH results, clinical follow-up, subsequent cytopathology or surgical specimens were reviewed. In two such cases, no further results were available; those cases were removed from analysis. Of the remaining 96 cases, there were 3 false negative and 4 false positive results. The sensitivity of CK20 staining was 87.5% and the specificity was 94.4%.
Conclusions: Based on 87.5% sensitivity and 94.4% specificity in this study, the use of CK20 as a marker of urothelial neoplasia has potential for practical use in the laboratory. CK20 is more widely available, cost effective, and has a faster turnaround time than FISH analysis. Potential drawbacks include varying levels of individual experience with stain interpretation, distraction by and misinterpretation of CK20 positive superficial urothelial cells, and inability to recognize atypical cell groups on stained slides.
Monday, February 28, 2011 1:00 PM
Poster Session II # 261, Monday Afternoon