Theory and practice of predictive assays in radiation therapy

An overview of the field of predictive assays is presented. It has been written with the many clinicians and scientists in mind who would like to become better acquainted with the general scope, principles and themes in the field.

Predictive assays have yielded much valuable information about the radiobiology of tumors e.g. the overall treatment time for rapidly proliferating tumors should be kept to a minimum. However, alteration of current treatment protocols based on results from predictive assays is still a matter of debate. What justification do we have to change established treatment protocols? A necessary and sufficient justification would be when the test value indicated an alternative treatment producing a better outcome, i.e. higher survival, improved local control, etc. Necessary but not always sufficient justification is correlation between the parameter measured and clinical outcome, if insufficient clinical benefit can be derived even when this is known. It is not sufficient that a test be demonstrated to be discriminatory. It must discriminate a sufficient number of patients, and its use must provide the patient with useful therapeutic alternatives. These parameters measured by predictive assays may well interact radiobiologically, and restricting observations to just one assay is probably insufficient for reliable indications. In the future, it is more likely that a panel of tests will be performed, and clinical decisions based on multi-parametric analysis of biopsy material.

In the following overview general predictive assay theory is presented followed by a brief introduction to some of the more established assays and finally some guidelines are suggested for the development of new assays.