The aggressiveness of brain cancer requires quantitative evaluation tools that can detect response to therapy early in order to guide treatment. Quantitative FDG-PET imaging has been used extensively, both within clinical trials at the Dana-Farber Cancer Institute (DFCI) and elsewhere, to evaluate the response of novel cancer therapies between a baseline (pre-treatment) and a follow-up (post-treatment) scan. The standard analysis approach, involving manual delineation of regions of interest, is robust but limited in scope, time-consuming, and subjective. In the context of neuroimaging, we propose to develop more objective methods that can indicate changes anywhere in the brain and correct for the confounding induced by global and regional changes in normal brain metabolism. This is achieved by a voxelwise comparison between the pre-treatment and post-treatment 3D scans, preceded by spatial registration, segmentation and background adjustment, and followed by significance thresholding. Summary measures of the generated voxelwise change maps are evaluated as predictors of survival in clinical trials. The methods in this proposal will provide a better tool for radiological assessment of patient progression or response and a better standard for evaluation of therapy in clinical trials. Capitalizing on complimentary expertise and interests in image analysis, this unique collaboration between the departments of Imaging and Biostatistics at DFCI holds the promise of developing fundamental methodologies that can immediately be evaluated and utilized in previous and future clinical trials.
The methods in this proposal will provide a better tool for radiological assessment of patient progression or response to treatment in brain cancer and a better standard for evaluation of therapy in clinical trials.
| Qin, Lei; Schwartzman, Armin; McCall, Keisha et al. (2017) Method for detecting voxelwise changes in fluorodeoxyglucose-positron emission tomography brain images via background adjustment in cancer clinical trials. J Med Imaging (Bellingham) 4:024006 |
| Naylor, Melissa G; Cardenas, Valerie A; Tosun, Duygu et al. (2014) Voxelwise multivariate analysis of multimodality magnetic resonance imaging. Hum Brain Mapp 35:831-46 |
| Guo, Mengye; Yap, Jeffrey T; Van den Abbeele, Annick D et al. (2014) Voxelwise single-subject analysis of imaging metabolic response to therapy in neuro-oncology. Stat 3:172-186 |
