Development and validation of a next generation sequencing based microsatellite instability assay for routine clinical use

Abstract

Colorectal cancer (CRC) is the second most common cancer in both men and women. Approximately 3-5% of CRCs show microsatellite instability (MSI) caused by germline defects in mismatch repair genes. In addition, 12% of sporadic CRCs show MSI. Currently, MSI is tested using a fragment analysis based assay not suitable for high throughput testing. Knowledge of microsatellite instability affects prognosis, surveillance and treatment of CRCs and MSI testing is now recommended for all newly diagnosed CRCs. As a result, development of high throughput approaches is desirable. The focus of my work was to develop and validate a high throughput sequence based MSI assay. Initially, I tested 25 (7-9bp) mononucleotide markers, previously identified from in silico analyses, using a cohort of 55 CRCs, and selected 8 markers which collectively could discriminate between MSI-high (MSI-H) and microsatellite stable (MSS) cases. To define the optimal parameters to discriminate between MSI-H and MSS samples, I tested these 8 markers and 9 long (8-12bp) mononucleotide markers identified in a parallel study, across a panel of 141 CRC samples. This allowed development of a scoring scheme for the 17 markers, which achieved 96% sensitivity and 100% specificity. I validated this scheme using an independent cohort of 70 CRCs without knowing their MSI status. The assay achieved a 100% sensitivity and specificity. Finally, I assessed the ability of short repeats to allow inference of the clonal variation within both FFPE (7) and fresh (4) MSI-H CRCs by analysing multiple samples from each cancer. I was able to infer the lineage relationship between primary tumour and lymph node metastasis in three cases and to construct phylogenetic trees for all cancers for which multiple samples were available illustrating the utility of these markers for understanding of CRC clonal variation.