The cancer state of a cell is characterized by alterations of important cellular processes such as cell proliferation, apoptosis, DNA-damage, repair, etc. Some of these alterations may result in modifications of the expression of genes that participate in the pathways involved in those processes. From this simple observation it follows that the expression of genes associated with cancer related pathways should exhibit differences between the normal and the transformed states. We explore various means to find those differences. Interestingly, these differences can only be identified when groups of genes, as opposed to isolated genes, are considered. In typical studies of cancer using gene expression arrays all the genes participating in a DNA array are used when comparing cohorts of cancer patients with control subjects. In our study only the genes associated with a specific pathway are considered. This results in a substantial reduction of dimensionality, thereby allowing for a considerable increase of the signal to noise of the analyses, while preserving the biological information contained in the pathways. We analyze 6 different pathways (p53, Ras, Brca, DNA damage repair, NFkb and b-catenin) and 4 different types of cancer: colon, pancreas, prostate and kidney. Our results recapitulate most of the existing knowledge of the involvement of those pathways in those cancers, suggesting that it is possible to identify pathways that are specifically involved in given cancers. Our analysis constitutes proof of principle that it may be possible to identify pathway involvement in cancers whose biology is poorly known, using gene expression data.