Abstract

Personalised cancer medicine based on the genetic milieu of individual tumours has long been postulated but until recently this concept was not supported by clinical evidence. Over the last five years, we have pioneered a research program that has translated into the first example of personalised treatment of colorectal cancer (CRC). Current treatment of this disease includes EGFR-targeted therapy with monoclonal antibodies (moAbs), but these agents provide significant clinical benefit in only about 10% of unselected patients. Using clinical samples and innovative cellular models we have shown that CRC harboring molecular alterations in downstream effectors of the EGFR (such as KRAS and BRAF) are resistant to treatment with the EGFR-targeted moAbs. These findings were confirmed by independent laboratories with such compelling evidence that they were rapidly translated into clinically applicable predictive biomarkers (genomic signatures) currently used to select patients for anti- EGFR therapy. Even when they are selected based on the molecular profile of their tumours only a fraction of patients with metastatic colorectal cancer (mCRC) have clinical benefit from therapy with anti-EGFR antibodies, which calls for the identification of novel biomarkers for better personalised medicine. Direct transfer xenografts of tumour surgical specimens conserve the interindividual diversity and the genetic heterogeneity typical of the tumours of origin, potentially combining the flexibility of preclinical analysis with the informative value of population-based studies. We have established cohorts of 200 distinct, genetically characterised mCRC ‘xenopatients' and exploited them to discover novel determinants of therapeutic response and new druggable driver oncoproteins. This knowledge was translated back into the clinic by profiling colorectal cancer patients who had received distinct targeted therapies.