Bevacizumab (Avastin™) is a monoclonal antibody against the protein vascular endothelial growth factor A (VEGF). VEGF is over-expressed in many tumour types, and compounds that inhibit this pathway can reduce tumour growth and sensitise tumours to other anti-cancer drugs. Bevacizumab has been tested in clinical trials in many solid tumour types, most often in combination therapy. Many of these showed that it can improve progression-free survival when combined with chemotherapy, but little effect has been observed on overall survival. However, variable responses indicate that it would be valuable to target therapy towards selected groups of patients, and the current Response Evaluation Criteria in Solid Tumours (RECIST) guidelines are proving inadequate to predict which patients should be so targeted.

Adrian Jubb and Adrian Harris from the Weatherhall Institute of Molecular Medicine, University of Oxford, UK have reviewed putative biomarkers for bevacizumab efficacy based on data from clinical trials. This wide-ranging review included radiological imaging, clinical, and molecular biomarkers. Imaging-based methods of investigating the activity of bevacizumab, which is directed to the tumour vasculature, studied have included computer tomography, MRI and ¹⁸F-fluorothymidine position emission tomography (FLT-PET) scanning. None of these methods were able to predict efficacy, but Jubb and Harris concluded that labelling bevacizumab to measure its binding to VEGF is a promising angle for continuing investigation.

VEGF binding to its receptor is known to cause vasodilation, and thus to decrease blood pressure. Thus, inhibition of VEGF binding is likely to increase blood pressure, and this has been observed clinically with hypertension occurring in 5-18% of patients treated with bevacizumab. The hypothesis that observed hypertension might be a good surrogate marker for bevacizumab-VEGF binding and thus for antibody efficacy has been tested in a number of clinical trials. Although the results have not been completely consistent, several studies did indicate a clear association between overall or progression free survival and hypertension. These studies did, however, raise questions. There was no consensus on which blood pressure levels should be recorded as hypertension, and there are, in any case, questions about using an adverse event as a biomarker. The clearest consensus was over the necessity of monitoring blood pressure levels carefully in patients taking bevacizumab.

Following evidence from pre-clinical studies, Jubb and Harris also considered tumour infiltration by myeloid cells as a possible marker for resistance to bevacizumab. They suggested that patients suffering from neutrophil depletion (neutropenia) during therapy might benefit from the addition of bevacizumab to their drug regimens, and further that therapy designed to prevent neutropenia might have a negative effect on bevacizumab efficacy. Few studies, however, have tested this correlation, and results have been ambiguous in those few.

Measuring the concentrations of proteins in the bloodstream is an attractive biomarker strategy, as it is cheap and easily accessible. The most obvious protein biomarker to select for testing bevacizumab efficacy is VEGF itself, but no studies yet have found changes in VEGF concentrations during therapy to have consistent predictive value. One study found circulating levels of the protein intercellular adhesion molecule 1 (ICAM-1) to be at least potentially predictive of survival benefit. Genetic polymorphisms, however, are already showing more promise as biomarkers. Studies in breast and colorectal cancer have shown two polymorphisms in the VEGF gene to be predictive of either overall or progression free survival under regimens including bevacizumab.  There is also evidence that polymorphisms in other genes in the VEGF pathway may have predictive value.

Jubb and Harris conclude that the only biomarkers that can currently be said to have any value in predicting bevacizumab efficacy are hypertension and VEGF polymorphisms. However, some other approaches, particularly imaging, are showing promise. Much more work will be needed to determine the most effective biomarkers and thus to select the cohorts of patients that are most likely to benefit from this therapy.

Article:  Jubb, A.M. and Harris, A.H. (2010). Biomarkers to predict the clinical efficacy of bevacizumab in cancer. Lancet Oncology 11, 1172-83.