Enabling non-invasive surveillance to measure and predict early cancer evolution

75% of UK cancer deaths are associated with late diagnosis, when our treatments are insufficiently effective. To diagnose cancer early we must regularly survey our tissues for signs of developing tumours to enable early, curative interventions. Effective surveillance will require detection of rare abnormal cell populations in a vast background of healthy cells and prediction of cancer risk. Hence we must (1) understand how abnormal cells progress towards cancer and (2) develop methods to overcome this ‘impurity’ in signals associated with cancer surveillance. We will undertake a survey of early cancer evolution using multi-sampled (pre)-malignant tissues from 12 organs and several thousand individuals. Selection, mutational processes and heterogeneity during early carcinogenesis will be associated with progression risk, germline-associated risk and environmental exposures. We will develop two technologies designed to monitor for high-risk clonal evolution by extracting signals of clonal composition (ECLIPSE2.0) or chromosomal instability (QUASAR) in highly impure (under 1%), non-invasively collected samples such as blood or the oesophageal capsule sponge. These technologies will be benchmarked and applied to clinical samples from the TRACERx study of early-stage non-small cell lung cancer and the BEST4 screening study for Barrett’s oesophagus. These results will inform the next generation of early detection assays.