Development of Novel Therapeutic Response Biomarkers

Abstract

Biomarkers that can indicate the best treatment option for each patient could greatly improve pancreatic cancer survival. Markers need to be practical to use in a timely fashion in order to change the choice of therapy. In vitro or ex vivo studies are useful in identifying potential markers, but these may not have relevance to marker profiles of in situ tumors, and adequate quality of tumor tissue may not be routinely available in patients with advanced disease, and so blood-based markers of systemic determinants of response may be more attrac-tive. Any marker, tissue-or blood-borne, needs to be tested in clinical studies involving multiple populations before entering routine use. These studies cannot rely just on prognosis as one individual's survival may be improved by therapy but still be significantly shorter than another whose survival was independent of therapy. Ideally an objective measure of response that links to survival benefit should be used to evaluate a biomarker. However, this may not be possible for adjuvant therapy where the tumor is removed before treatment begins and the link between survival and response in an advanced setting is not always reliable. Survival on its own is a poor surrogate for response, and its use may lead to confusion of prognostic and response markers unless used within large clinical trials. Adverse responses to treatment such as rash linked to survival may be an alternative measure. Difficulties in defining the level of beneficial response make empirical identification of response biomarkers difficult. Theory-based studies have more power to identify and validate markers, but the determinants of drug response are complex, and popular (but potentially misguided) beliefs about 1 specific proteins may lead to multiple testing and hence type 1 errors. Grouping biomolecules (proteins, RNA, metabolites, or DNA sequences) into marker panels linked to function, for example, grouping proteins that determine mesen-chymal transition of cancer cells or which define the nature of stroma, may offer a way forward. Alternatively, functional analysis alone, including level of immune response, may allow the most beneficial therapy to be directed to each patient.