Programmatic Theme: Translational Bioinformatics

Abstract: Increased cell density (CD) correlates with tumor infiltration and malignancy in gliomas. Estimating CD non-invasively could help guide diagnosis and treatment for patients. We collected extensive pre-operative magnetic resonance imaging (MRI) including anatomic, diffusion, and dynamic contrast sequence. Using local intensity information from MRI and CD measurements from biopsies, we trained random forest model to predict CD. The random forest estimated CD with R²=0.73, showing great predictability of clinically relevant pathology.

Learning Objective: Be able to describe how advanced MR imaging techniques correlate with underlying tumor cellularity and state which techniques are most informative for predicting cellularity.

Authors:

Evan Gates (Presenter)
University of Texas MD Anderson Cancer Center

Jonathan Lin, University of Texas MD Anderson Cancer Center
Jeffrey Weinberg, University of Texas MD Anderson Cancer Center
Jackson Hamilton, University of Texas MD Anderson Cancer Center
Sujit Prabhu, University of Texas MD Anderson Cancer Center
John Hazle, University of Texas MD Anderson Cancer Center
Gregory Fuller, University of Texas MD Anderson Cancer Center
Veera Baladandayuthapani, University of Texas MD Anderson Cancer Center
David Fuentes, University of Texas MD Anderson Cancer Center
Dawid Schellingerhout, University of Texas MD Anderson Cancer Center