Designing personalised cancer treatments

The concept of personalised medicine for cancer is not new. It arguably began with the attempts by Salmon and Hamburger to produce a viable cellular chemosensitivity assay in the 1970s, and continues to this day. While clonogenic assays soon fell out of favour due to their high failure rate, other cellular assays fared better and although they have not entered widespread clinical practice, they have proved to be very useful research tools. For instance, the ATP-based chemosensitivity assay was developed in the early 1990s and is highly standardised. It has proved useful for evaluating new drugs and combinations, and in recent years has been used to understand the molecular basis of drug resistance and sensitivity to anti-cancer drugs. Recent developments allow unparalleled genotyping and phenotyping of tumours, providing a plethora of targets for the development of new cancer treatments. However, validation of such targets and new agents to permit translation to the clinic remains difficult. There has been one major disappointment in that cell lines, though useful, do not often reflect the behaviour of their parent cancers with sufficient fidelity to be useful. Low passage cell lines - either in culture or xenografts are being used to overcome some of these issues, but have several problems of their own. Primary cell culture remains useful, but large tumours are likely to receive neo-adjuvant treatment before removal and that limits the tumour types that can be studied. The development of new treatments remains difficult and prediction of the clinical efficacy of new treatments from pre-clinical data is as hard as ever. One lesson has certainly been that one cannot buck the biology - and that understanding the genome alone is not sufficient to guarantee success. Nowhere has this been more evident than in the development of EGFR inhibitors. Despite overexpression of EGFR by many tumour types, only those with activating EGFR mutations and an inability to circumvent EGFR blockade have proved susceptible to treatment. The challenge is how to use advanced molecular understanding with limited cellular assay information to improve both drug development and the design of companion diagnostics to guide their use. This has the capacity to remove much of the guesswork from the process and should improve success rates.