A65
Development of expression based biomarkers in NSCLC: a study of intratumour teterogeneity using FFPE tissue
Robert Holt1, Dean Fennell2, Ian Paul2, Kieran McManus2, Paul Kelly4, Peter Kerr1, Timothy Davison1, Vitali Proutski1, Richard Kennedy1, Jacqueline James2
1Almac Diagnostics, Craigavon, UK, 2CCRCB, Queens University, Belfast, UK, 3Royal Victoria Hospital, Belfast, UK, 4The Belfast Health and Social Care Trust, Belfast, UK
There is considerable interest in the use of gene expression data to generate prognostic and predictive biomarkers from non-small cell lung cancer (NSCLC) tissue. Since individual NSCLC tumours exhibit histological diversity, it is possible that different areas of individual tumours will exhibit distinct molecular profiles. This represents a potential problem for the development of expression-based biomarkers.
To date, intratumoural gene expression has primarily been studied to determine the feasibility of using small biopsies in expression studies, e.g. in ovarian cancer. NSCLC studies describing EGFR mutations suggest that these tumours demonstrate tissue heterogeneity, however gene expression data is limited.
We are currently developing an early NSCLC prognostic gene signature using approximately 1500 FFPE tumour samples. The study of intratumoural gene expression in these samples will play a crucial role in facilitating the development of a robust gene signature. Ten samples were selected, representing five adenocarcinoma and five squamous carcinoma tumours. Five tissue blocks representing different parts of each tumour were analysed. RNA was extracted from both macrodissected and whole FFPE sections and processed onto the Lung Cancer DSA, a lung cancer specific DNA microarray platform.
Bias and variations in gene expression were assessed at the intra- and inter patient (FFPE block) level and between adenocarcinoma and squamous samples. Unsupervised clustering and visualisation of gene expression data indicated that variations in gene expression were present between FFPE blocks taken from a single patient. However, the variations in gene expression observed between individual patients and disease states were greater. Gene expression data was also fitted to a repeated measures ANOVA model to assess the statistical significance of observed variations within and between samples and disease states. The implications for the development of expression-based biomarkers are discussed.
