Abstract P2-02-07: Predicting Response to Bevacizumab Therapy in Primary Breast Cancer Using Dynamic Contrast-Enhanced Magnetic Resonance Imaging

Abstract

Abstract Background: Anti-angiogenic therapy holds much promise for the treatment of breast cancer. In practice however, only a subset of patients who receive these drugs demonstrate a significant response to therapy. A key challenge therefore is to elucidate markers that are predictive of response to anti-angiogenic agents such as bevacizumab, and which would enable the selection of patients who would get the most benefit from these expensive therapies. Materials and Methods: We used high temporal resolution dynamic contrast-enhanced magnetic resonance imaging (DCE-MRI) to assess tumor vascularity in 20 patients with primary breast cancer. Patients were imaged both before and two weeks after single dose Bevacizumab therapy (15mg/kg). Pharmacokinetic modelling techniques were used to quantify the volume transfer constant Ktrans, the rate constant kep, and the fractional volume of the extra-vascular extracellular space ve. Specifically, we used Tofts model with a population-based arterial input function (modified Fritz-Hansen) to model the contrast agent concentration time courses on a voxel-wise basis. Non-enhancing voxels were detected automatically with the use of a Bayesian noise model, and the corresponding pharmacokinetic parameter values for these voxels were set to zero. The median pharmacokinetic parameter values over the tumor volumes of interest were then computed both pre-and post Bevacizumab. Results: We found marked variation across patients in the baseline level and percentage change in median Ktrans, kep and ve following Bevacizumab. In particular, median Ktrans at baseline ranged form 0.12 to 0.88. Changes in median Ktrans varied from −97% to +19% across all patients, with an average change of −49%. Notably, we found a highly significant negative correlation (r = −0.92, P = 1e-08) between the absolute change in median Ktrans and the median Ktrans at baseline. In particular, tumors with a high median Ktrans at baseline demonstrated the greatest change in Ktrans following Bevacizumab therapy, whereas tumors with low median Ktrans at baseline demonstrated relatively little change in Ktrans. Discussion: Although Ktrans is a complex function of vessel permeability, surface area, and tumor blood flow, it has previously been demonstrated to be a reliable biomarker of response to anti-angiogenic therapy in a number of different cancers. Our results illustrate that therapy-induced changes in Ktrans can be predicted from the value of Ktrans at baseline, and hence DCE-MRI scans may enable the selection of primary breast cancer patients who show the greatest response to single-dose Bevacizumab therapy. Whether this will translate into longer term benefit and improvements in outcome for patients remains to be shown. The relationship between baseline and pre-/post-therapy change in Ktrans with the corresponding changes in gene expression is currently under study in a larger number of patients. Figure available in online version. Citation Information: Cancer Res 2010;70(24 Suppl):Abstract nr P2-02-07.