Abstract

Requirements for Clinical Utility of Expression Signatures and SNP Profiles based on Microarray Data

Bertrand Jordan, Emeritus Research Director, National Centre for Scientific Research

DNA microarrays have allowed extensive expression profiling of normal and pathological samples as well as whole-genome association studies aimed at finding the genes involved in frequent, multigenic diseases. While this has provided much new knowledge, the transition to actual clinical applications has been limited. Expression profiles have been applied for prognostic and predictive purposes, particularly in oncology and for breast cancer. Many tests (>50) have been proposed and published, but only a few are actually used and only one or two can be considered commercially successful. This is due to the need for a very robust implementation, to the required demonstration of analytical validity, clinical validity and above all clinical utility, in addition to regulatory requirements and cost/reimbursement issues. Whole-genome SNP profiling has revolutionised the genetics of complex diseases, with the successful implementation of GWAS studies that have at last provided solid identification of loci and genes influencing common, multigenic diseases. However the relative risks associated with the deleterious alleles identified are generally small, of the order of 1.2 to 1.5, and their predictive value at the level of an individual is accordingly very limited. Profiles marketed direct to consumers have also met with regulatory problems. As additional knowledge accumulates in the future, these analyses will acquire more definite medical value.