ACRIN 6684: Assessment of tumor hypoxia in newly diagnosed GBM using FMISO PET and MRI.

person Elizabeth Robins Gerstner Massachusetts General Hospital, Boston, MA info_outline Elizabeth Robins Gerstner, Zheng Zhang, James R Fink, Mark Muzi, Lucy Hanna, Erin Greco, Akiva Mintz, Lale Kostakoglu, Edward A Eikman, Melissa Prah, Kathleen M Schmainda, Gregory A Sorensen, Daniel Barboriak, David A. Mankoff

Authors person Elizabeth Robins Gerstner Massachusetts General Hospital, Boston, MA info_outline Elizabeth Robins Gerstner, Zheng Zhang, James R Fink, Mark Muzi, Lucy Hanna, Erin Greco, Akiva Mintz, Lale Kostakoglu, Edward A Eikman, Melissa Prah, Kathleen M Schmainda, Gregory A Sorensen, Daniel Barboriak, David A. Mankoff Organizations Massachusetts General Hospital, Boston, MA, Brown University, Providence, RI, University of Washington, Seattle, WA, Wake Forest University, Winston-Salem, NC, Icahn School of Medcn At Mount Sinai, New York, NY, H. Lee Moffitt Cancer Center & Research Institute, Tampa, FL, Medical College of Wisconsin, Milwaukee, WI, Duke University, Durham, NC, University of Pennsylvania, Philadelphia, PA Abstract Disclosures Research Funding NIH Background : Tumor hypoxia is a potent mediator of treatment resistance because of the negative impact on the efficacy of radiation and associated poor tumor delivery of cytotoxic chemotherapy. We designed a multicenter study to evaluate if pre-treatment tumor hypoxia as measured by FMISO PET and abnormal tumor vascularity as assessed by MRI are associated with worse survival. Methods : Patients with newly diagnosed GBM who had residual enhancing or nonenhancing tumor by local MRI read after surgery were eligible. Patients were treated with standard RT and temozolomide alone or with an additional investigational agent. Prior to start of chemoRT, a PET scan was acquired 110min after FMISO injection and 3 blood draws were acquired to allow for calculation of tumor to blood ratios. MRIs were also performed and included standard anatomic sequences as well as perfusion (DSC) and permeability (DCE) imaging. Cox models tested the association of PET and MRI parameters with OS (primary endpoint) and PFS (secondary endpoint). ROC curve analyses examined whether these parameters could predict survival at 1 year. Results : Fifty patients from 11 centers were enrolled. Thirty-eight, 37 and 31 patients had evaluable FMISO PET scans, DSC imaging and DCE imaging, respectively. Median age was 57 (range 27-77). In the univariate model, higher peak SUV was associated with worse survival (HR 1.54, 95%CI 1.00-2.36, p = 0.048) as was mean Ktrans (HR 1.17, 95%CI 1.02-1.34, p = 0.024) and median Ktrans (HR 1.32, 95%CI 1.01-1.72, p = 0.045). After adjusting for age and enhancing tumor volume, only mean Ktrans remained statistically significant. Higher median Ktransas well as normalized rCBV and rCBF were associated with shorter PFS. Max SUV, peak SUV, standardized rCBV and normalized rCBF were predictive of survival at 1 year. Conclusions : Higher baseline tumor hypoxia was associated with shorter survival, re-iterating the poor prognostic implications of tumor hypoxia. Increased vascular permeability and perfusion measurements at baseline were also associated with worse survival. These results raise the possibility that hypoxia may mediate treatment resistance in areas of tumor expected to have adequate delivery of intravascular therapies across the BBB. Clinical trial information: NCT00902577

Clinical