Laila Adel Mohsen MBBCh, MSC, Abstract Co-Author: Nothing to Disclose

Adam M.H. Young, Abstract Co-Author: Nothing to Disclose

Veronica Shi, Abstract Co-Author: Nothing to Disclose

Owen Michael Thomas MBBCh, DPhil, Abstract Co-Author: Nothing to Disclose

Victoria Lupson, Abstract Co-Author: Nothing to Disclose

Nasim Sheikh Bahaei MD,FRCR, Presenter: Nothing to Disclose

Amy Frary, Abstract Co-Author: Nothing to Disclose

Rajesh Jena MD, FRCR, Abstract Co-Author: Nothing to Disclose

Jonathan Harvey Gillard MD, Abstract Co-Author: Consultant, GlaxoSmithKline

Stephen J. Price, Abstract Co-Author: Nothing to Disclose

The EORTC/NCIC treatment protocol had improved survival of glioblastoma (GBM) patients. Still patients die from progressive disease. This is because tumor cells in the invasive margin are not identified with conventional imaging. This study uses multimodal MRI to non-invasively assess the microenvironment of the invasive region of GBM. Our aim was to try to confirm that the occult invasive margin of GBM, as identified by diffusion tensor imaging (DTI), does differ from other peri-tumoral regions.

Thirty patients with GBM were imaged pre-operatively at 3T. The imaging protocol consists of standard anatomical, standard DTI, gradient echo perfusion and multivoxel short echo time (TE: 35 msec) MR spectroscopy (MRS) sequences. The isotropic (p) and anisotropic (q) components of the diffusion tensor were calculated as previously described1. Comparing these parameter maps, peritumoral invasion was identified as those regions that demonstrated an increase in the isotropic component (p) but that had normal anisotropic diffusion (q). These regions were then used to select appropriate MRS voxels of interest for analysis. The normalized regional cerebral blood volume (rCBV) was calculated. The metabolite concentrations and rCBV were compared to normal-appearing adjacent and contralateral white matter. MRS was analysed using LCModel. A paired t-test was used.

The invasive margin had significantly higher rCBV than the control peritumoral regions (p=0.001). Seven metabolites demonstrated significant concentration differences in the invasive region compared to normal-appearing white matter. The invasive margin demonstrated a reduction in N-Acetylaspartate (NAA; p=0.001), Myo-inositol (Ins; p=0.001), Creatine (Cr; p<0.0001) and the glutamate/glutamine cycle (Glx; p=0.001). Comparisons between normal-appearing adjacent and contralateral white matter demonstrated no difference for MRS but not for rCBV, which was higher in the normal appearing peritumoral regions as well.

The microenvironment of invasive regions shows higher tumor activity than other peritumoral regions. This method allows noninvasive assessment of the invasive behavior of this margin.

DTI can idenitfy the invasive margins of GBM. This method may be used to tailor the radiotherapy field according to the patient’s tumor invasive pattern.

Mohsen, L, Young, A, Shi, V, Thomas, O, Lupson, V, Sheikh Bahaei, N, Frary, A, Jena, R, Gillard, J, Price, S, Non-invasive Assessment of Tumour Microenvironment in Invasive Margins of Glioblastomas: A Multimodal Imaging Pilot Study.  Radiological Society of North America 2012 Scientific Assembly and Annual Meeting, November 25 - November 30, 2012 ,Chicago IL. http://archive.rsna.org/2012/12025776.html