Prognostic value of spatial heterogeneity in FMISO-PET images of hypoxia in brain cancer patients

449 Objectives In malignant brain tumors, larger FMISO T/Bmax (maximum tissue uptake to blood) values, which indicate the most hypoxic elements in a tumor, are associated with poorer clinical outcomes. We used a measure of spatial heterogeneity, developed in the context of FDG imaging of sarcoma, to investigate if the distribution of hypoxia in the tumor mass can be quantified to further differentiate the risk of progression and death. Methods We evaluated heterogeneity in 40 brain cancer patients who had FMISO PET studies with concurrent blood sampling as part of RT planning and clinical follow up. The patients were scanned between their diagnostic surgery/biopsy and before RT/ChX. Sixteen patients had follow-up FMISO scans following completion of RT/ChX. Registered MR images acquired within days of PET scanning were used to guide spatial assessment. We used a spatial heterogeneity measure constructed from a tubular representation of the tumor mass and a simplified radial analysis of uptake transverse to the tubular axis [O’Sullivan 2011]. An open-source software implementation was used to apply this method to FMISO tumor uptake data. A multivariate Cox regression analysis was used to assess the added prognostic benefit of heterogeneity beyond T/Bmax and other established predictors of clinical outcome in the patient cohort. Results Multivariate analysis showed that heterogeneity was an independent predictor of disease progression and death. More heterogeneous tumors are associated with significantly (p<0.05) shorter TTP and survival. Conclusions Quantification of the spatial heterogeneity of hypoxia in brain tumors adds prognostic value. References : O'Sullivan, F. et al. A Statistical Modeling Approach to the Analysis of Spatial Patterns of FDG-PET Uptake in Human Sarcoma. IEEE Trans. Med. Imaging 30:2059-2071, 2011. Research Support Supported by NIH CA042045 and RR17229, and Science Foundation Ireland grant PI 11/1027.