30 Chapter 2 The use of tissue microarrays in biomarker screening During the production of tissue microarrays (TMAs) (18, 19), a tissue core (0.6 to 3 mm diameter) is sampled from an archival formalin-fixed paraffin-embedded donor block and repositioned in a paraffin acceptor block. Subsequent sectioning of the acceptor block, results in slides representing tens to hundreds of cancer specimens. As such, immunohistochemical staining of TMA sections allows rapid screening of the behavior of biomarkers across large sample population (20). Additional advantages are sparse use of resection specimens, incorporation of internal controls, a decrease in staining variability (no batch effects), possibly automated scoring, lowering in costs, and facilitation of multi-center collaborations (20). Furthermore, when properly annotated, the TMA technology allows examination of the association between the biomarker of interest and patient, tumor, and/or histopathological characteristics, facilitating the identification of prognostic, predictive and therapeutic targets in large independent patient cohort assuring sufficient statistical power (20). Despite the many benefits of using TMAs, the technique also has it downsides (20). Most of them are similar to those involving the use of whole tissue sections (WTSs), such as tissue quality, use of reliable antibodies, standardization of methodologies, antigen loss, and more specifically production and cutting of the actual arrays (20). Besides these practical pitfalls, a major concern in the use of TMAs is spatial tumor heterogeneity (20). For instance, despite hematoxylin and eosin based selection of tumor areas by the pathologist, deeper cuts may not contain any tumor cells, which especially is of concern when analyzing in situ lesions (20). Of major concern is the examination of biomarkers that are heterogeneously expressed across a cancer. Examination of the correlation between homogeneously expressed biomarkers (Pearson r > 0.6) in a breast cancers setting, led to the general consensus that two cores with 0.6 mm diameter are representative for WTSs (20, 21). However, for heterogeneously expressed biomarkers the general notion is that this number should be considerably higher (22, 23). Therefore, it has also been proposed to construct TMAs in such a way that different tumor areas are represented in separate cores, allowing assessment of intratumoral heterogeneity (20). Finally others have argued that for large TMA studies sampling error is compensated by sample size, therefore including only one core per tumor (24).
RkJQdWJsaXNoZXIy MTk4NDMw