RESUMEN
BACKGROUND: The advantage of tissue microarray (TMA) is its ability to efficiently analyze large numbers of tissue specimens in a methodologically uniform way. The reliability of TMAs, especially with regard to clinicopathological characterizations, when compared to conventional immunohistochemistry (IHC) was evaluated. MATERIALS AND METHODS: Seventy-two embedded tissue sections from lung cancer specimens were stained with monoclonal antibodies against the tumor-associated markers TA-MUC1 and Lewis Y. Three representative cores of every tumor were embedded in a paraffin array multiblock. The IHC was evaluated by the immunoreactive score (IRS). RESULTS: The data for the TMA IHC and the conventional IHC were concordant (kappa > or = 80%) for both markers. Likewise, discordance (McNemar's test) was low, and sensitivity and specificity were above 80% for both markers. In the samples with high positive expression, the concordance increased (kappa > or = 90%), discordance disappeared (McNemar p = 1.0), and sensitivity and specificity increased above 90% for both markers. Using Cox regression models, all the clinicopathological dependencies were equivalent for both techniques and both markers. CONCLUSION: Immunohistochemistry with tissue microarrays is valid and provides results equivalent to conventional immunohistochemistry with respect to expression patterns and clinicopathological characterizations.