We conclude that: (1) BAP1 immunohistochemistry is relatively insensitive in the context of sarcomatous and desmoplastic mesotheliomas, but as a matter of time and cost efficiency may nonetheless be a useful first approach to the problem; (2) deletion of p16 by FISH is considerably more sensitive, but there remain a proportion of cases in which p16 is not deleted; (3) a small improvement in sensitivity can be achieved by using both markers; (4) in the context of a spindle cell malignant tumor in the pleura or peritoneum, which morphologically might be a metastatic sarcomatoid carcinoma or a mesothelioma, the finding of BAP1 loss favors mesothelioma, but p16 FISH cannot be used to separate sarcomatous mesotheliomas from sarcomatoid carcinomas.
Copy number analyses by molecular inversion probe assays of the three spindle cell carcinoma samples revealed little overall genomic CNAs with only minor changes identified (fraction of the genome altered; 1.3%-6.4%), but with a common 9p21.3 loss in 2 out of 3 samples, with CDKN2A (p16) being a likely candidate.