An Analysis of Spatial Patterns of PET-FDG Uptake in Human Sarcoma
Finbarr O'Sullivan1, Eric Wolsztynski1, Janet N. O'Sullivan1, Todd Richards2, Ernest C. Conrad2, Janet F. Eary2
1Statistics, University College Cork, Cork, Ireland; 2Center for Orthopedic and Sports Medicine and Division of Nuclear Medicine, University of Washington, Seattle, WA, United States

Clinical experience with PET-FDG scanning of sarcoma has established spatial heterogeneity in the FDG utilization pattern within the tumor mass as a key prognostic indicator of patient survival. But it may be that a more detailed assessment of the tumor FDG utilization pattern would provide additional insights into patient risk. The present work develops a statistical model for this purpose. The approach is based on a tubular representation of the tumor mass with a simplified radial analysis of uptake patterns transverse to the tubular axis. This technique provides novel ways of characterizing the overall profile of the tumor and introduces a potential approach for the measurement of its phase characteristics. A series of PET-FDG studies from 185 patients is used to formally evaluate the prognostic benefit. Significant improvements in the prediction of patient survival and progression are obtained by including a core-phase measure generated from the tumor profiling analysis. This core-phase measure identifies central low metabolism tissues which can arise from necrosis and fluid, fat or cartilage accumulations. The work confirms that more detailed quantitative assessments of the spatial pattern of PET imaging data of tumor masses may provide improved prognostic information for potential input to patient treatment decisions.

Keywords: Positron emission tomography; Spatial statistics; Tumor profiling; Phase assessment

Biography: Dr O'Sullivan works on statistical methodologies for spatial and temporal analysis of diagnostic imaging data. He collaborates with a clinically orientated multi-disciplinary PET imaging program based at the University of Washington. The focus of this research is on the metabolic imaging of cancer and its response to therapy. Dr O'Sullivan has developed novel statistical approaches to the assessment of cancer heterogeneity and the non-parametric analysis of radio-tracer kinetics.