Cyclin D1 immunohistochemical staining to separate benign from malignant mesothelial proliferations.

ABSTRACT

The separation of benign from malignant mesothelial proliferations is a morphologically difficult problem. Mutations/deletions of components of the Hippo pathway are frequent in malignant mesotheliomas, and one downstream effect of aberrant Hippo signaling is increased production of cyclin D1. We examined expression of cyclin D1 nuclear staining in two tissue microarrays containing 52 reactive epithelial mesothelial proliferations, 51 reactive spindle cell mesothelial proliferations, 54 epithelial mesotheliomas, and 22 sarcomatous/desmoplastic mesotheliomas. When present, cyclin D1 staining was always strong, hence the arrays were scored as 0, 1-25%, 26-50%, 51-75%, and 76-100% staining. Both arrays showed a similar pattern. Reactive epithelial proliferations generally showed no staining (42/52 cases) or 1-25% staining (10/52 cases) with no cases showing >25% staining. Overall for reactive epithelial proliferations the maximum staining was 14.8% and mean 1.1 ± 2.9%. For epithelial mesotheliomas 39/54 (72%) cases demonstrated >25% staining, with 8/54 in the 26-50% staining range, 9/54 in the 51-75% range, and 22/54 in the >75% range. Combinations of staining using cyclin D1 >50% plus BAP1 or MTAP loss in epithelial mesotheliomas produced about a 10% increase in sensitivity. Reactive spindle cell proliferations showed a broader range of staining with 27/51 in the 1-25% range, 5/51 in the 26-50% range, and 1/51 >50%. Eleven of 22 sarcomatous/desmoplastic mesotheliomas scored 50% or greater. We conclude that for epithelial mesothelial proliferations, the finding of >50% of tumor cells staining supports a diagnosis of epithelial mesothelioma with 100% specificity but only modest (57%) sensitivity.