RNA biomarkers in colorectal cancer

Colorectal cancer (CRC) develops and progresses through a systematic selection for (epi) genetic alterations that drive the transformation from normal colon epithelium to adenocarcinoma. These changes affect both noncoding RNAs and mRNAs and so define the clinical behaviour of cancer cells within a distinctive host genetic and environmental context. Although earlier diagnosis and more effective treatment modalities have decreased mortality from CRC, prognostic stratification and adjuvant therapy selection after surgery remain dependent on broad descriptive classifications, opportune histological markers of poor prognosis and chemotherapy efficacy data derived from diverse CRC populations. Crucially, there is significant inter- and intra-individual variability in response to, and tolerance of, chemotherapy treatments. These limitations explain the small clinical benefit of new agents studied in contemporary phase III trials. Molecular assays have the potential to address these constraints and there has been intense interest in the identification of clinically relevant molecular biomarkers. These must be easy to obtain and quantify and ideally represent steps in well-understood carcinogenic pathways or host-response mechanisms. Although some biomarkers can provide broad prognostic information based on CRC subtype (e.g. MSI status) or can somewhat predict response to targeted therapies (e.g. KRAS), no RNA-based biomarkers have entered routine clinical practice. This is due, in part, to the genetic heterogeneity of both patients and CRC. In addition, serious underlying issues with regards to study design, poor technical protocols, inadequate quality controls and inappropriate data analysis prevent successful translation of research results. Consequently, the identification of clinically relevant panels of biomarkers will depend not just on further advances in our understanding of CRC biology, but will need to be coupled with appropriate study designs and more suitable, standardised and transparent techniques.