Advanced Imaging of Brain Tumors

Room 701 A


Chairs: Michael Alan Kraut and Meng Law


Prog #

13:30 716. 31P MRS of Human Brain Tumors at 3T Using 1H –31P Polarization Transfer

Jannie P. Wijnen1, Dennis WJ Klomp1, Arend Heerschap1

1Radboud University Nijmegen Medical Centre, Nijmegen, Netherlands

Although 31P MRS of tumors in humans provides more relevant information on phospholipid metabolism than 1H MRS it is hampered by sensitivity limitations. In this study we demonstrate the feasibility of an optimized localized 1H-31P polarisation transfer method to examine phosphorylated compounds in brain tumors with increased sensitivity and resolution at 3T. The changes observed in tumor tissue for ratios of individual phosphomonoester and diester compounds may be valuable biomarkers for tumor diagnosis and treatment follow up.

13:42 717. Multi-Exponential Diffusion Imaging: Clinically Feasible Multiple B-Value Diffusion Imaging for High Grade Gliomas

Inas S. Khayal1, 2, Colleen McGue1, Soonmee Cha1, Susan M. Chang1, Roland G. Henry1, Sarah J. Nelson1

1University of California, San Francisco, San Francisco, California , USA

The goal of this study was to develop and validate a two b-value diffusion sequence capturing the deviation from mono-exponential decay signal (dS) in a time efficient clinical scan; and to test how this signal can differentiate tumor (contrast enhancing lesion) from edema (non-enhancing lesion) as compared to the conventional apparent diffusion coefficient (ADC) and fractional anisotropy (FA) in high grade gliomas. This technique shows spatial variation within enhancing and non-enhancing lesions which may be of interest for tumor characterization and may be utilized in distinguishing progressive tumor from treatment response. Further studies will verify this variation with image-guided biopsy.

13:54 718. The Functional Diffusion Map (FDM) as an Early Imaging Biomarker for High-Grade Glioma: Correlation with Conventional Radiologic Response and Overall Survival

Craig J. Galbán1, Daniel A. Hamstra1, Charles R. Meyer1, Timothy D. Johnson1, Pia C. Sundgren1, Christina Tsien1, Theodore S. Lawrence1, Larry Junck1, David J. Ross1, Alnawaz Rehemtulla1, Brian D. Ross1, Thomas L. Chenevert1

1University of Michigan, Ann Arbor, Michigan, USA

We ascertained if diffusion MRI, by functional diffusion mapping (fDM), could directly predict patient survival. Sixty patients with high grade glioma undergoing radiation therapy were enrolled in a study of intra-treatment MRI 1, 3, and 10 weeks after the initiation of treatment. Compared to conventional neuro-imaging, fDM provided an earlier assessment of equal predictive value. While the combination of fDM and radiological response provided a more accurate prediction of patient survival than either metric alone.

14:06  719. A Novel Membrane-Permeant Contrast Agent for Ultra-Sensitive Brain Tumor Detection by MRI

Mohanraja Kumar1, Zdravka Medarova1, Pamela Pantazopoulos1, Anna Moore1

1Massachusetts General Hospital, Charlestown, USA

One of the key challenges hindering effective therapy against brain cancer is defined by the inability to detect brain tumors at an early stage. With this in mind, we designed a novel membrane-permeant contrast agent, MN-MPAP, which consists of a superparamagnetic iron oxide nanoparticles, for MR imaging, labeled with the near-infrared dye Cy5.5, for near-infrared optical imaging, and conjugated to myristoylated polyarginine peptides (MPAP), as a membrane translocation module. In vivo MRI established the value of MN-Cy5.5-MPAP for tumor localization and delineation. T2 map analysis demonstrated higher accumulation of MN-Cy5.5-MPAP in tumors than control MN-Cy5.5, as reflected by the change in tumoral T2 relaxation times following injection of the contrast agent.

14:18 720. Glioma Detection Maximized Through Intracellular and Extracellular MR-Guided Optical Fluorescence Tomography

Summer Gibbs-Strauss1, Julia A. O'Hara, Scott C. Davis1, Brian William Pogue1

1Dartmouth College, Hanover, New Hampshire, USA

MR-guided optical fluorescence spectroscopy of glioma brain tumors was studied, contrasting the detectability of the 9L tumor with the U-251 tumor.  While the 9L tumor was delinated by T1 Gd-contrast MR, the U-251 was not as well resolved.  Intracellular and extracellular optical contrast agents however showed significantly better delineation of the U-251 tumor though, due to increased metabolic function.

14:30 721. Improved Diffusion MR Fiber Tracking for Neurosurgical Applications

Jeffrey I. Berman1, Danny J. Clark1, Mitchel S. Berger1, Roland G. Henry1

1University of California San Francisco, San Francisco, California , USA

In this study, a residual bootstrap probabilistic q-ball fiber tracking technique was used to delineate the motor tract in four brain tumor patients. Results of probabilistic and deterministic q-ball fiber tracking are compared to the currently used DTI fiber tracking technique.  Q-ball fiber tracking methods were able to delineate the motor tract to lateral portions of the motor cortex while DTI fiber tracking could not traverse regions of crossing white matter fibers. The HARDI acquisition used for this study was acquired in a clinically feasible timeframe and q-ball fiber tracking has the potential to be routinely used for surgical planning.

14:42 722. Random-Walk Based Tractography Simulation for the Growth of Brain Tumors

Poe-Jou Chen1, 2, Wei-Ting Zhang2, Rakesh K. Jain3, T T. Batchelor3, A Gregory Sorensen2

1Massachusetts Institute of Technology, Cambridge, Massachusetts, USA; 2Athinoula A. Martinos Center for Biomedical Imaging, Charlestown, Massachusetts, USA; 3Massachusetts General Hospital, Charlestown, Massachusetts, USA

A random walk based tractography model is proposed to simulate the growth of glioblostams multiforma (GBMs). The simulated tumor evolution was compared with the real growth observed with two MR images of a patient acquired with one month difference. Results demonstrate the potential of this random walk model and the proposed model will be evaluated on more patient data to prove its feasibility in real clinical applications

14:54 723. Differentiation of Low-Grade Glioma Subtypes Using Mutiparametric MR Data

Wei Bian1, Inas S. Khayal, 12, Colleen McGue1, Suja Saraswathy1, Soonmee Cha, Susan M. Chang1, Sarah J. Nelson1

1University of California San Francisco, San Francisco, California , USA; 2UCSF/UCB Joint Graduate Group in Bioengineering, San Francisco, California , USA

To differentiate low-grade glioma (LGG) subtypes, forty five newly diagnosed grade II glioma patients (17 oligodendroglioma, 11 astrocytoma, and 17 oligoastrocytoma) underwent proton MR spectroscopy imaging (MRSI), diffusion weighted imaging (DWI), and perfusion weighted imaging (PWI). Normalized metabolite intensities were estimated from MRSI, apparent diffusion coefficient (ADC) from DWI, and relative cerebral blood volume (rCBV) from PWI. Discriminant analysis was performed to determine the parameters that better differentiate LGG subtypes. Median ADC and 25th percentile rCBV helped differentiate oligodendroglioma from astrocytoma, and median ADC and Choline helped differentiate oligodendroglioma from oligoastrocytoma. Better classification accuracy was achieved when using these mutiparametric MR data.

15:06 724. Monitoring Brain Tumor Response to Radiation by Sodum MR Imaging

Keith Thulborn1, Fred Damen, Aiming Lu1, Ian Atkinson1, John Villano1, Howard Halpern, Bulent Aydogan2, Michael Flannery1, Hagai Ganin1

1Univ. of Illinois at Chicago, Chicago, Illinois, USA; 2Univ. of Chicago, Chicago, Illinois, USA

Quantitative sodium imaging can be used to follow brain tumor response to radiation during initial treatment. A lack of response predicts treatment failure and its location. Such data allows modification of the initial treatment and the possibility to achieve an improved outcome for these patients.

15:18 725. Which is Best for Classifying Brain Tumours from 1H Spectra: Expert Interpretation, Metabolite Ratios, Classifiers or the INTERPRET Decision-Support System?

Margarida Juliŕ-Sapé1, 2, Indira E. Coronel2, Carles Majós3, Ana Paula Candiota1, 2, Marta Serrallonga3, 4, Monica Cos3, Carles Aguilera3, Juan José Acebes5, John R. Griffiths6, Carles Arús, 12

1CIBER-BBN, Centro de Investigación Biomédica en Red en bioingenieria, Biomateriales y Nanomedicina, Cerdanyola del Valles, Spain; 2Universitat Autonoma de Barcelona, Cerdanyola del Valles, Spain; 3CSU de Bellvitge, L’Hospitalet de Llobregat, Spain; 4Telemedicine Clinic, Barcelona, Spain; 5Hospital Prínceps d’Espanya, CSU de Bellvitge, L’Hospitalet de Llobregat, Spain; 6Cancer Research UK Cambridge Research Institute, Li Ka Shing Centre, Cambridge, UK

Four methods for analysis of MR spectra were tested to determine which method worked best. Data from 40 consecutive adult patients with abnormal brain masses were prospectively analyzed. Formal criteria for MRS evaluation were used: a) Structured description of the MRS by expert spectroscopists; b) Analysis with the INTERPRET decision-support system; c) Calculation of a mIno/Gly ratio; d) Classifiers at short and at long TE. The best system was the INTERPRET DSS. It performed better than expert evaluation, as its embedded classifier incorporates quantitative measurements, whereas expert interpretation is by nature, qualitative.