With the increasing use of antiangiogenic agents for the treatment of cancers, establishing biomarkers of response and resistance has become a priority for oncologists and pharmaceutical companies. This urgency comes from the need to select the patients most likely to benefit from these high-cost therapies. It also stems from the necessity of identifying new targets to prevent the invariable escape from these therapies, which target specifically or primarily the vascular endothelial growth factor (VEGF) pathway. But the much needed biomarkers remain elusive. One of the reasons is the still unclear mechanism(s) of action of these drugs. Blocking VEGF can have antivascular and normalizing effects on the tumor vasculature, which may not necessarily translate into clinical responses as evaluated by criteria based on tumor size measurements such as the Response Evaluation Criteria in Solid Tumors. Moreover, excessive antivascular effects (when using high doses) might induce a transient response, but could lead to severe toxicities, as well as to more aggressive tumors, as seen in mouse models. Similarly, vascular normalization alone (with no cytotoxic treatment) might not be sufficient to shrink tumors or halt their growth, as demonstrated in mice. Finally, some of the actions of antiangiogenic agents could be systemic. For example, antiangiogenic agents could affect trafficking and function of hematopoietic progenitor cells and effector immune cells. This could result in promotion or delay in tumor growth depending on the hematopoietic cell type involved. Given this complexity, it is most likely that for each cancer and each agent, we might need a specific set of biomarkers for good prediction, and that these biomarkers will be mechanism specific. Ideally, these biomarkers should be relatively easy to measure by imaging or in bodily fluids using standardized protocols. For plasma or serum biomarkers, this could be achieved readily, given the multiple and reliable options to measure various proteins. To date, only a few randomized trials have retrospectively integrated circulating biomarker evaluations, and unfortunately, none have yet identified a valid circulating biomarker candidate. In non–small-cell lung cancer (NSCLC), the anti-VEGF antibody bevacizumab was shown to be effective when combined with chemotherapy in two randomized phase III trials. In a phase II/III study of bevacizumab with chemotherapy in NSCLC patients, a high baseline circulating plasma VEGF level did not predict patient survival, despite correlating with the response evaluated by Response Evaluation Criteria in Solid Tumors. Similarly, baseline soluble intracellular adhesion molecule 1 (sICAM1) was an independent prognostic factor of overall survival in patients treated with chemotherapy with chemotherapy alone or with bevacizumab. No correlation was seen with other intuitive biomarker candidates, such as basic fibroblast growth factor or sE-Selectin. Therefore, identifying biomarker candidates for prospective evaluation in randomized antiangiogenic trials remains an outstanding challenge in NSCLC and other cancers. The comprehensive biomarker study by Hanrahan et al published in this issue of Journal of Clinical Oncology is a step in the right direction. These investigators explored a set of 35 plasma biomarkers in NSCLC patients at four time-points after antiangiogenic therapy alone with the VEGF receptor 2 tyrosine kinase inhibitor (TKI) vandetanib, chemotherapy alone, or a combination of the two. Vandetanib monotherapy transiently increased the levels of circulating interleukin 8 (IL-8; at day 8) and VEGF (at day 43), and decreased the levels of its soluble receptor VEGF receptor 2 (sVEGFR2; at day 43). In contrast, chemotherapy alone did not change circulating VEGF or IL-8, but transiently decreased sVEGFR2, IL–1 receptor antagonist, matrix metalloproteinase 9, and IL-12, and increased monocyte chemotactic protein-1 plasma levels at day 8. Surprisingly, vandetanib with chemotherapy did not significantly change circulating VEGF, sVEGFR2 or IL-8 levels, but transiently decreased IL-12 and matrix metalloproteinase 9, and increased monocyte chemotactic protein-1 in plasma at day 8. Hanrahan et al also explored possible correlations between the changes in these biomarkers after treatment and the outcome in individual patients. They report that lower levels of sICAM1 at day 8 after treatment were significantly associated with poorer treatment outcome in the groups of patients who received vandetanib. Although exploratory in nature and with a relatively modest sample size, Hanrahan et al report data from a randomized, threearm trial. The study has important implications. First, it confirms that pursuing mechanistic biomarkers (ie, circulating proteins with known proor antiangiogenic activity) shortly after treatment initiation might be a fruitful approach, as many of the biomarker changes occur rapidly after the onset of therapy. Consistent with this, we have shown JOURNAL OF CLINICAL ONCOLOGY E D I T O R I A L VOLUME 28 NUMBER 2 JANUARY 1
Read full abstract
Jun 1, 2013
Isamu Okamoto + 2
Open Access
