Use of desmin immunohistochemistry to distinguish between mesothelial cells and carcinoma in serous fluid cell block preparations

OBJECTIVE: To evaluate the utility of immunohistochemical stains for desmin in discriminating mesothelial cells from adenocarcinoma in serous fluid cell block preparations. STUDY DESIGN: Cell block preparations from 22 cases (representing 18 patients) that were positive for carcinoma and 5 cases that were negative for malignancy were immunostained with an antibody to desmin. Positive staining was evaluated and scored semiquantitatively in both tumor cells and background mesothelial cells in the malignant cases and mesothelial cells in the negative controls. Staining was evaluated with a score of 0-3 for intensity and 0-5 for distribution. The sum of the two scores was recorded as the total score (TS). RESULTS: Mesothelial cells from all the carcinoma and benign cases stained with desmin (median TS=3D5.5, range 4-8), typically strong in intensity and widespread in distribution. Positivity was observed in carcinoma cells in all cases, typically weak and focal (range 2-4). Using a total score of 4 as a cutoff for definitively positive staining, desmin staining was positive in mesothelial cells in 25/25 cases and carcinoma cells in 1/22 cases (P<.0001, Fisher's exact test). Additionally, using the Mann-Whitney ranked sum test on the 20 cases with evaluable mesothelial cells, the medians of the total scores for mesothelial cells (5.5) and carcinoma cells (2.5) were significantly different (P<.0001). CONCLUSION: A total score of 4 was significantly associated with mesothelial cell staining. Use of desmin immunohistochemical staining in cell block preparations may be helpful in distinguishing between mesothelial cells and carcinoma