The immunohistochemical diagnosis of mesothelioma is commonly made by using a battery of antibodies that reacts with lung adenocarcinomas but not with epithelial mesotheliomas. Only recently have markers that are often expressed in mesotheliomas but not in adenocarcinomas been recognized. Some of these markers, however, require frozen tissue sections, whereas others are not commercially available, or their value remains controversial. In a recent publication, it was suggested that immunostaining for cytokeratin 5/6 could assist in distinguishing epithelial mesothelioma from lung adenocarcinoma. To determine the practical value of cytokeratin 5/6 immunostaining in the diagnosis of mesothelioma, 40 formalin-fixed, paraffin-embedded epithelial pleural mesotheliomas, 30 pulmonary adenocarcinomas, 93 nonpulmonary adenocarcinomas, 15 squamous carcinomas of the lung, 5 large cell undifferentiated carcinomas of the lung, and 12 metastatic transitional cell carcinomas to the lung were stained with the same antibody, which was obtained from a commercial source. Cytokeratin 5/6 reactivity was observed in all 40 mesotheliomas, but there was none in any of the 30 pulmonary adenocarcinomas. Focal or weak reactivity was observed in 14 of 93 nonpulmonary adenocarcinomas (10 of 30 ovarian, 2 of 10 endometrial, 1 of 18 breast, I of 7 thyroid, 0 of 10 kidney, 0 of 10 colonic, and 0 of 8 prostatic). All 15 squamous carcinomas of the lung, 6 of 12 transitional cell carcinomas metastatic to the lung, and 3 of 5 large cell undifferentiated carcinomas of the lung expressed cytokeratin 5/6. It is concluded that cytokeratin 5/6 immunostaining is not only useful in separating epithelial pleural mesotheliomas from pulmonary adenocarcinomas but also can assist in distinguishing epithelial mesotheliomas from nonpulmonary adenocarcinomas metastatic to the pleura.