In most validation studies of tissue microarrays (TMAs), a fixed number of cores with a given diameter are analyzed to determine the degree of accuracy by which the TMA represents the whole section. The statistical model described in the present study predicts this property for various combinations of 2 core sizes (0.6 mm and 1.2 mm) and different core numbers. The model was based on artificial TMA core biopsies generated from Ki-67 and active caspase-3 immunostains of 40 canine lymphoma samples. Positivity was scored on a continuous scale, and a large number of cells were analyzed with the help of semiautomated cell counting. Despite considerable differences in range and distribution of Ki-67 and active caspase-3 positivity values, the model predictions showed a high degree of agreement for both markers. Comparison of 0.6 mm and 1.2 mm cores indicated that the use of small cores necessitates inclusion of a larger number of samples but requires counting a markedly smaller number of cells. Suitability of TMAs to determine the immunophenotype of the whole section was assessed using 2 different combinations of core sizes and numbers. Both displayed a high degree of concordance with the whole section (kappa(0.6) = 0.79; kappa(1.2) = 0.91). The present study provides a basis for the use of TMAs in future high-throughput immunohistochemical investigations of selected markers in canine lymphomas. The statistical model presented can be used to determine an optimal TMA design depending on a desired accuracy.