Statistical Points to Consider in the Precision Testing of Immunohistochemistry Assays
View Presentation View Presentation
*Gene Anthony Pennello, Food and Drug Administration 

Keywords: reproducibility, CLSI EP5-A, variance components, categorical data, multinomial

Immunohistochemistry (IHC) assays are designed to detect and visualize immunological reactions in tissues and cells by use of microscopy. In the precision testing of IHC assays, challenges include (i) heterogeneity in a tissue specimen, (ii) limited number of sections per tissue specimen, (iii) pathologist reader variability in histopathological evaluation, and (iv) summarizing imprecision of semi-quantitative test results. I present designs of precision experiments investigating multiple sources of variation (e.g., lab, reagent lot, day, run) when only a limited number of sections are available per tissue sample and when tissue heterogeneity is expected. For semi-quantitative tests, I present summary tables describing imprecision and propose the Gini index as a summary measure. To address reader variability, average positive and negative percent agreements are considered.