Glioblastoma (GBM, WHO grade IV) is the most common and highest-grade astrocytoma with uniformly dismal prognosis. In contrast to the lower grade gliomas (grades II and III), GBMs present radial growth rates almost 10 times as fast. Three pathophysiologic features, including necrosis, cellular pseudopalisades surrounding necrosis, and microvascular hyperplasia, are considered as hallmarks for GBM prognosis and are believed to be relevant to the drastically accelerated disease progression. Additionally, a population of Glioma Stem Cells (GCSs) has been found pertinent to the new neovascular formation, leading to GBM tumor outward growth. However, the definitive roles of these pathologic features and the associated signaling networks in GBM progression and their interactive mechanisms remain ill defined, because of limited capacity of biomarker observation and data analysis. The proposed project will create a new avenue for GBM research by leveraging computational power and machine intelligence for analyses of large-scale pathology images and spatially referenced, multiplexed biomarker data. A comprehensive system and analytical infrastructure will be developed to quantitatively investigate 1) spatial distributions f GBM invasion-related pathologic structures and GSCs, 2) functions of the in-situ signaling genes and regulatory networks responsible for GBM progression, and 3) the significance of morphologic, molecular, and signaling network variation across GBM microenvironments. A set of scalable image processing algorithms for analyses of spatial distributions and co-localizations of phenotypic features and biomarkers will be developed and applied to whole-slide pathology brightfield and quantum dot immunohistochemistry images of GBMs. Derived features and boundaries of phenotypic structures, and spatial distributions of biomarkers will be archived and integrated in a database for pathology and in-situ molecular imaging data where scientific queries can be invoked to support the analysis of GBM tumor progression mechanisms. In particular, multiplexed quantum dot immunohistochemistry (mQD-IHC), a new biomarker staining technique, will be leveraged to show the locations and expressions of multiple biomarkers of interest within the same tissue spatial reference. The combined expertise in large-scale pathology image process, high performance data computation, mQD-IHC staining technique for multiplexed biomarker investigation, and phenotype- genotype data integration supported by high-throughput data query power from customized database enables a unique research vehicle to investigate GBM tumor expansion mechanisms. Notably, the proposed research work can be generalized to catalyze other computational cancer research involving large-scale data and phenotype-genotype information integration. Therefore, it will present a broad impact on future translational research.

Public Health Relevance

Integrated analysis of spatial configurations of tissue pathologic features and underlying molecular biomarkers accounting for cancer progression is an important way to understand tumor expansion mechanisms. To advance in this research field, I propose to develop automated image analysis tools, database-based spatial integration methods, and mathematical models for tumor microenvironment representation to support quantitative investigations on Glioblastoma (GBM) brain tumor invasion mechanisms with definitive pathology imaging and multiplexed molecular biomarker data from whole-slide images of GBM specimens.

Agency
National Institute of Health (NIH)
Institute
National Cancer Institute (NCI)
Type
Mentored Quantitative Research Career Development Award (K25)
Project #
5K25CA181503-04
Application #
9315122
Study Section
Special Emphasis Panel (ZCA1)
Program Officer
Jakowlew, Sonia B
Project Start
2014-09-01
Project End
2019-08-31
Budget Start
2017-09-01
Budget End
2018-08-31
Support Year
4
Fiscal Year
2017
Total Cost
Indirect Cost
Name
Emory University
Department
Biomedical Engineering
Type
Schools of Medicine
DUNS #
066469933
City
Atlanta
State
GA
Country
United States
Zip Code
30322
Zhang, Tuo; Kong, Jun; Jing, Ke et al. (2018) Optimization of macaque brain DMRI connectome by neuron tracing and myelin stain data. Comput Med Imaging Graph 69:9-20
Gomes, Jeremias; de Melo, Alba C M A; Kong, Jun et al. (2018) Cooperative and out-of-core execution of the irregular wavefront propagation pattern on hybrid machines with Intel? Xeon Phi™. Concurr Comput 30:
Mukherjee, Subhas; Tucker-Burden, Carol; Kaissi, Emily et al. (2018) CDK5 Inhibition Resolves PKA/cAMP-Independent Activation of CREB1 Signaling in Glioma Stem Cells. Cell Rep 23:1651-1664
Ross, James L; Cooper, Lee A D; Kong, Jun et al. (2017) 5-Aminolevulinic Acid Guided Sampling of Glioblastoma Microenvironments Identifies Pro-Survival Signaling at Infiltrative Margins. Sci Rep 7:15593
Barreiros Jr, Willian; Teodoro, George; Kurc, Tahsin et al. (2017) Parallel and Efficient Sensitivity Analysis of Microscopy Image Segmentation Workflows in Hybrid Systems. Proc IEEE Int Conf Clust Comput 2017:25-35
Zhang, Pengyue; Wang, Fusheng; Teodoro, George et al. (2017) AUTOMATED LEVEL SET SEGMENTATION OF HISTOPATHOLOGIC CELLS WITH SPARSE SHAPE PRIOR SUPPORT AND DYNAMIC OCCLUSION CONSTRAINT. Proc IEEE Int Symp Biomed Imaging 2017:718-722
Teodoro, George; Kurç, Tahsin M; Taveira, Luís F R et al. (2017) Algorithm sensitivity analysis and parameter tuning for tissue image segmentation pipelines. Bioinformatics 33:1064-1072
Rossetti, Blair J; Wang, Fusheng; Zhang, Pengyue et al. (2017) DYNAMIC REGISTRATION FOR GIGAPIXEL SERIAL WHOLE SLIDE IMAGES. Proc IEEE Int Symp Biomed Imaging 2017:424-428
Teodoro, George; Kurc, Tahsin; Andrade, Guilherme et al. (2017) Application Performance Analysis and Efficient Execution on Systems with multi-core CPUs, GPUs and MICs: A Case Study with Microscopy Image Analysis. Int J High Perform Comput Appl 31:32-51
Zhang, Changming; Mukherjee, Subhas; Tucker-Burden, Carol et al. (2017) TRIM8 regulates stemness in glioblastoma through PIAS3-STAT3. Mol Oncol 11:280-294

Showing the most recent 10 out of 23 publications