Bevacizumab-Induced Diffusion Restriction in Patients With Glioma: Tumor Progression or Surrogate Marker of Hypoxia?

Abstract

TO THE EDITOR: Therapy with bevacizumab results in high response rates in patients with malignant glioma. However, the mechanisms of action of bevacizumab are poorly understood. Blood-brain barrier reconstitution and vascular normalization have been proposed as phenoma distinct from classical effects of antiangiogenic therapy (ie, induction of hypoxia). Refined magnetic resonance imaging may help to clarify this issue. We, therefore, read with great interest the case report recently presented by Gerstner et al in Journal of Clinical Oncology that described an area of restricted diffusion in a patient with malignant glioma treated with bevacizumab as analyzed by diffusion weighted imaging (DWI). The diffusion-restricted area was biopsied. Histology displayed diffuse tumor progression. Therefore, the authors conclude that diffusion-restricted areas in patients with glioma who are treated with antiangiogenic therapy could represent a special pattern of recurrent tumor. We agree with the authors that diffusion restriction occurs in patients treated with bevacizumab. In our recently published analysis, 18 patients treated with bevacizumab were prospectively analyzed for areas of diffusion restriction on DWI, and apparent diffusion coefficient maps were generated. We found that diffusion-restricted areas accompanied by a corresponding apparent diffusion coefficient decrease occurred in 13 patients (65%). Magnetic resonance imaging–based perfusion analysis performed in two patients showed reduction of regional cerebral blood flow and volume that corresponded with diffusion restriction during treatment with bevacizumab. In one patient, a diffusion-restricted area was biopsied, and histology demonstrated atypical necrosis. Furthermore, immunohistochemistry showed heavy nuclear staining for hypoxia inducible factor-1 . Thus, although we agree that increases in DWI might reflect tumor recurrence where dense cell packing occurs, such as in lymphoma, our analysis suggests that alternative causes for diffusion restriction exist. In our series, DWI restriction was common in patients whose diseases were responding to treatment. The atypical necrosis together with the hypoxia inducible factor-1 induction would be compatible with chronic hypoxia induced by prolonged antiangiogenic therapy. Whether antiangiogenic therapy truely induces hypoxia might be relevant given the therapy-resistant phenotype and diffuse infiltrating growth pattern often observed in gliomas after antiangiogenic therapy, as hypoxia might contribute to both phenoma.