Cytopathologic differential diagnosis of malignant mesothelioma, adenocarcinoma and reactive mesothelial cells: A logistic regression analysis

Diagnostic Cytopathology. 2009 Jan;37(1):4-10. [Link]

Cakir E, Demirag F, Aydin M, Unsal E.

Department of Pathology, Ankara Ataturk Chest Diseases and Chest Surgery Education and Research Hospital, Ankara, Turkey. arabaci.ebru@gmail.com

Abstract

Distinguishing malignant mesothelioma, adenocarcinoma and reactive mesothelial proliferation in both cytologic and surgical pathologic specimens is often a diagnostic challenge. Conventional cytomorphologic assessment is an important step in the differential diagnosis of these entities.
The pleural effusion cytologies from 40 cases of malignant mesothelioma, 40 cases of adenocarcinoma and 30 cases of reactive mesothelial proliferation diagnosed between 1997 and 2007 were reviewed. Twenty-seven cytologic features which are regarded as useful in the differential diagnosis of mesothelioma, adenocarcinoma and benign mesothelial proliferation were assessed. These cytologic features were subjected to a stepwise logistic regression analysis. Three features were selected to distinguish malignant mesothelioma from adenocarcinoma: giant atypical mesothelial cell (P = 0.0001), nuclear pleomorphism (P = 0.0001) and acinar structures (P = 0.0001), the latter two being characteristics of adenocarcinoma. The variables selected to differentiate malignant mesothelioma from reactive mesothelial cells were: cell ball formation (P = 0.0001), cell in cell engulfment (P = 0.0001) and monolayer cell groups (P = 0.0001), the latter being a feature of benign mesothelial proliferation. When these selected variables were subjected to a stepwise logistic regression analysis, the logistic model correctly predicted 90% of cases of benign mesothelial proliferation versus 97.5% of malignant mesothelioma and 92.5% of malignant mesothelioma versus 92.5% of adenocarcinoma.

Keywords: mesothelioma, adenocarcinoma, mesothelial proliferation, cytodiagnosis