Digital pathology for the validation of tissue microarrays in peripheral T-cell lymphomas

Systematic validation of construction and analysis parameters when using tissue microarray (TMA) in rare, morphologically heterogenous entities such as peripheral T-cell lymphoma (PTCL) is not reported. We describe a tissue-saving virtual TMA to predetermine the number of cores needed to represent whole tissue sections (WTS) from the same biopsies, using automated and traditional manual methods for the quantification of immunohistochemical stains. Whole paraffin hematoxylin and eosin- and immunohistochemical (CD2, CD30, and Ki-67)-stained sections from 30 PTCLs were digitalized. A virtual TMA with six 1-mm cores per slide was designed to compare agreements in the immunohistochemical scoring. Using digital image analysis and manual stereological counting, immunohistochemical positivity was quantified. Associations were analyzed using the Bland-Altman and correlation plots. In PTCL, we report that 4 cores are required to represent WTS results (ie, agreement within ±10%). High concordance was demonstrated between digital results obtained with WTS compared with 4-core virtual TMA (correlation coefficients: 0.89-0.98), and in the comparative evaluation of 4-core virtual TMA by digital image analysis versus manual stereology (correlation coefficients: 0.91 to 0.99). Virtual TMAs provide an efficient tool for optimizing and validating TMA construction parameters when planning a study. The method can be applied to the same tissues used in a subsequent formal study, without wasting scarce tissue resources. In PTCL, TMAs constructed with four 1-mm cores are representative of WTS. In parallel tests using TMAs and WTS from PTCLs, there is a high level of agreement comparing automated digital with manual stereological methods for the quantification of immunohistochemical biomarker staining.