Richman, Susan D., Adams, Richard, Quirke, Phil, Butler, Rachel, Hemmings, Gemma, Chambers, Phil, Roberts, Helen, James, Michelle D., Wozniak, Sue, Bathia, Riya, Pugh, Cheryl, Maughan, Timothy, and Jasani, Bharat
Introduction Molecular characterisation of tumours is increasing personalisation of cancer therapy, tailored to an individual and their cancer. FOCUS4 is a molecularly stratified clinical trial for patients with advanced colorectal cancer. During an initial 16-week period of standard first-line chemotherapy, tumour tissue will undergo several molecular assays, with the results used for cohort allocation, then randomisation. Laboratories in Leeds and Cardiff will perform the molecular testing. The results of a rigorous pre-trial inter-laboratory analytical validation are presented and discussed. Methods Wales Cancer Bank supplied FFPE tumour blocks from 97 mCRC patients with consent for use in further research. Both laboratories processed each sample according to an agreed definitive FOCUS4 laboratory protocol, reporting results directly to the MRC Trial Management Group for independent crossreferencing. Results Pyrosequencing analysis of mutation status at KRAS codons12/13/61/146, NRAS codons12/13/61, BRAF codon600 and PIK3CA codons542/545/546/1047, generated highly concordant results. Two samples gave discrepant results; in one a PIK3CA mutation was detected only in Leeds, and in the other, a PIK3CA mutation was only detected in Cardiff. pTEN and mismatch repair (MMR) protein expression was assessed by immunohistochemistry (IHC) resulting in 6/97 discordant results for pTEN and 5/388 for MMR, resolved upon joint review. Tumour heterogeneity was likely responsible for pyrosequencing discrepancies. The presence of signet-ring cells, necrosis, mucin, edgeeffects and over-counterstaining influenced IHC discrepancies. Conclusions Pre-trial assay analytical validation is essential to ensure appropriate selection of patients for targeted therapies. This is feasible for both mutation testing and immunohistochemical assays and must be built into the workup of such trials. [ABSTRACT FROM AUTHOR]