Søren M Bentzen1, Francesca M Buffa, George D Wilson. 1. Department of Human Oncology, University of Wisconsin School of Medicine and Public Health, K4/316 Clinical Science Center, 600 Highland Avenue, Madison, WI 53792, USA. bentzen@humonc.wisc.edu
Abstract
INTRODUCTION: Tissue microarrays (TMAs) facilitate high-throughput immunohistochemical analysis of potential predictive and prognostic biomarkers in tumor or normal tissue samples. This technology and the practical issues involved in designing TMAs for translational research are reviewed. A main field of application of TMAs is in the search for predictive and prognostic markers in specific types of cancer. DISCUSSION: Standard data analytical approaches are discussed and some of the issues in applying these in practice are described. CONCLUSIONS: TMAs allow the collection of information-rich datasets on the simultaneous expression of multiple biomarkers. Supervised and unsupervised strategies have been developed for handling datasets where the number of covariates, in this case biomarker expression data, is large in relation to the number of patients in the sample. Some future research pathways are briefly presented together with the recent attempts to improve the reporting quality of biomarker studies.
INTRODUCTION: Tissue microarrays (TMAs) facilitate high-throughput immunohistochemical analysis of potential predictive and prognostic biomarkers in tumor or normal tissue samples. This technology and the practical issues involved in designing TMAs for translational research are reviewed. A main field of application of TMAs is in the search for predictive and prognostic markers in specific types of cancer. DISCUSSION: Standard data analytical approaches are discussed and some of the issues in applying these in practice are described. CONCLUSIONS:TMAs allow the collection of information-rich datasets on the simultaneous expression of multiple biomarkers. Supervised and unsupervised strategies have been developed for handling datasets where the number of covariates, in this case biomarker expression data, is large in relation to the number of patients in the sample. Some future research pathways are briefly presented together with the recent attempts to improve the reporting quality of biomarker studies.
Authors: Søren M Bentzen; Matthew Parliament; Joseph O Deasy; Adam Dicker; Walter J Curran; Jacqueline P Williams; Barry S Rosenstein Journal: Int J Radiat Oncol Biol Phys Date: 2010-03-01 Impact factor: 7.038