Quantitative immunochemistry in tissue sections : With special reference to urinary bladder carcinoma

Sammanfattning: Urinary bladder carcinoma affects about 2 000 people in Sweden every year. It is a heterogeneous disease and for adequate treatment decisions, prognostic tools are needed. Many of these are based on immunohistochemistry (MC) and hampered by poorly reproducible methodologies and subjective evaluations.This investigation aimed to evaluate and define the conditions for an objective and reproducible quantification of(MC) applied to routinely processed (formalin-fixed and paraffin-embedded) tissue material. To establish theseconditions, proliferation-related proteins, tumour-suppressor proteins and endothelial cell markers were studiedThree computerised colour-based image analysis methods, aiming to reduce user interaction, were developed and evaluated for variability and reproducibilty in quantification of IHC applied to histological sections. Themethods proved stable to variations in focus- and light-settings but revealed a sensitivity to poor IHC-qualityand choice of chromogen. Implementation of an external control system, used as base for computing theclassifier, further increased the stability.Using these methods, effects of varying tissue-handling and immunostaining conditions were analysed. Storageof both paraffin-blocks and -sections was deleterious to immunoreactivity but could be compensated for bytayloring of the heat induced epitope retrieval (HIER) protocol.A control system for MC based on cultured cells was developed to reflect variations in tissue-handling and IHC-conditions. Control and biopsy material showed similar sensitivity to section storage, MC- and HIER-protocolvariations. Furthermore, controls reflected variations in fixation time and proved stable to variations in sectionthickness.