Molecular Medicine, Gene-Expression Profiling and Molecular Diagnostics: Putting the Cart Before the Horse

Abstract

The terms ‘molecular medicine’ and ‘-omics technologies’ encompass a range of procedures that are promising to help delineate more precisely pathophysiological mechanisms, identify individuals at risk for disease and suggest novel targets for drug treatment. The enhanced consistency and ease of use of core technologies such as real-time PCR (qPCR) and microarrays are leading to the incorporation of molecular analyses into clinical practice [1]; the introduction of next-generation sequencing technology [2], which permits massively parallel sequencing of millions of sequences at once, is likely to see an acceleration of this trend. This is particularly true in cancer research, where there is expectant acknowledgment that epigenetic, genetic and gene-expression profiling of primary tumors, cancers and/or circulating nucleic acids and tumor cells is a helpful and critical new weapon in the perpetual endeavor to advance and finetune our clinical diagnostic potential [3]. For example, there have been extensive attempts aimed at identifying genetic markers and metastasis-associated mRNA signatures in solid tumors, and using them as prognostic indicators of distant cancer recurrence [4]. Other common applications aim to identify mRNA signatures associated with cancer risk or track changes in mRNA levels linked with response to therapy [5]. Consequently, there have been scores of publications reporting an abundance of supposed individual marker genes or diverse gene lists that are claimed to be associated with and even predictive of patient survival [6–9]. Unfortunately, very few of these genes or gene lists agree with one another and it is striking that, several years after metastasis-associated gene lists were first published in a number of solid cancers, there is still