Ramon Barajas MD, Presenter: Nothing to Disclose

John-Paul Jaewoon Yu MD, PhD, Abstract Co-Author: Nothing to Disclose

Christopher Paul Hess MD, PhD, Abstract Co-Author: Research support, General Electric Company

Cornelius Von Morze, Abstract Co-Author: Nothing to Disclose

Joanna Phillips, Abstract Co-Author: Nothing to Disclose

Soonmee Cha MD, Abstract Co-Author: Nothing to Disclose

Track-density imaging (TDI) allows for visualization of fine anatomical detail at submillimeter spatial resolution. The purpose of our study was to determine whether aggressive growth patterns of Glioblastoma (GBM), as assessed by histopathological analysis, can be quantified using TDI.

Eighteen patients with treatment-naïve GBM were prospectively recruited prior to surgical resection. Preoperative MRI on a 3T MR scanner was performed with standard anatomic sequences in addition to high angular resolution diffusion imaging (HARDI; TR/TE=9s/94 ms, b=2000 s/mm2, 55 directions, slice thickness 2.2 mm) reconstructed at 250 micron resolution using probabilistic streamline tractography with constrained spherical deconvolution (model order = 8, 0.1 mm step size, 1 million seed points) allowing for the production of track density (TD) and fractional anisotropy (FA) maps. Forty three tissue samples were collected utilizing stereotactic neurosurgical techniques. Histologic tissue analysis characterized aggressive features of GBM (tumor presence or absence, hypoxia, axonal architecture destruction, and cellular proliferation) using a four tier scale. TD and FA maps were co-aligned with anatomic images allowing for measurements from each biopsy site and contralateral brain producing relative values (rTD and rFA). Logistic regression assessed for differential rTD and rFA expression between biopsy sites based on low and high aggressiveness patterns. Linear regression tested for significance between imaging variables and histologic parameters.

Mean rTD was elevated within biopsy regions expressing increased levels of GBM aggressiveness compared to regions of low aggressive cellular features (P< 0.05). Mean rTD demonstrated a statistically significant positive correlation with cellular features of GBM aggressiveness (R> 0.28, P< 0.04). No significant differential expression or correlation was observed between rFA and histologic measures.

TDI provides submillimeter spatial resolution of white matter integrity and can quantify cellular features of GBM aggressiveness suggesting TDI provides unique spatial resolution of white matter tracts that is not achieved by conventional FA maps.

Correlation of rTD with cellular features of GBM aggression suggests TDI may be used as a robust measure of tumor infiltration.  Further study in a larger cohort of GBM patientsis currently underway.

Barajas, R, Yu, J, Hess, C, Von Morze, C, Phillips, J, Cha, S, Super-Resolution White Matter Track-Density Imaging Correlates with in Vivo Histopathologic Features of Glioblastoma Aggressiveness.  Radiological Society of North America 2011 Scientific Assembly and Annual Meeting, November 26 - December 2, 2011 ,Chicago IL. http://archive.rsna.org/2011/11010736.html