Aims: Tissue microarray (TMA) technology has been utilized for assessment of cancers including malignantpleural mesothelioma (MPM). Given the intralesional heterogeneity of MPM, it is questionable if TMAs can adequately represent MPMs. We here investigate the validity of TMAs for MPM. Methods: TMAs were constructed from at least five cores for each of 80 archival tumoursProcessed by two centres between 1994 and 2009. Thepercentage of cases correctly subtyped on TMAs compared with whole sections, in relation to the number of cores analysed, was calculated. Immunohistochemical labelling for calretinin and D2-40 wasperformed on TMAs and whole sections. To evaluate the validity of quantitative immunohistochemistry,percentages ofpositive cells were recorded and two-way analysis of variance (ANOVA)performed. Results: Five cores were assessable for 91% ofpatients. Four cores were sufficient to reach concordance with the whole-section result in 98% of cases for calretinin and 99% for D2-40. The correlation of the quantitative scores between the whole section and TMA cores was statistically significant (D2-40, rho=0.84,p<2.2e-16; calretinin, rho=0.65,p=7.9e-11). Neither the origin nor age of the blocks affected the results. Conclusion: If a minimum of four cores is used, TMA is an appropriate method for immunohistochemistry in MPM.