Search

In this study, we analyzed the MR spectroscopic and MR diffusion weighted imaging parameter differences among the subtypes of grade 3 brain tumor patients (anaplastic astrocytoma, oligodendroglioma, and oligoastrocytoma). We observed higher lipid values in the tumor regions of oligodendrogliomas than both astrocytomas and oligoastrocytomas. According to these results, there were not significant MR spectroscopic or MR diffusion differences between the subtypes of grade 3 brain tumors for this patient population and except lipid it would be acceptable to combine the data of these subgroups for future studies.

Susceptibility-weighted magnetic resonance imaging is a powerful tool for high resolution imaging of the vasculature, aiding in the diagnosis of many pathologic conditions. The technique is especially beneficial at higher field strengths where traditional sequences that measure cerebral blood volume suffer from severe distortions, rendering them inapplicable at 7T. However, conventional susceptibility- weighted imaging (SWI) sequences involve long scan times, on the order of 10 minutes for a 2 cm slab of coverage. This work implemented two partially parallel imaging reconstruction methods, 1) an autocalibrating parallel technique based on GRAPPA algorithm, and 2) sensitivity encoding (SENSE) for accelerating SWI of the brain at 7 Tesla. By employing two- fold under-sampling in the phase-encoding direction for both techniques, a two-fold reduction in scan time was simulated. Analysis of contrast ratios in large and small vessels compared to the surrounding brain parenchyma showed close agreement between the full and GRAPPA reconstructed datasets for both vessel sizes, while a decrease in the small vessel contrast was observed with SENSE.

Most methods that fit gamma-variate functions to perfusion curves suffer from convergence problems and lengthy processing times. In this paper, a technique that robustly and feasibly fits perfusion curves with a simplified form of the gamma-variate function is presented. The performance of the technique, as measured by its sensitivity to noise and tracer recirculation, was assessed by Monte Carlo methods with simulated curves. The algorithm was also applied to perfusion curves from 20 glioma patients and evaluated based on the fitting error during the first-pass phase. This technique may elucidate the clinical role of perfusion-weighted imaging in the future.

Lactate is a sensitive indicator of anaerobic glycolysis and reduced cellular oxygenation in living tissues. Mobile lipid resonances have been correlated with necrosis and cell breakdown. Reliable estimates of levels of lactate and lipid are of interest for evaluating prognosis and hence guiding the clinical management of patients with brain tumors. The goal of this study was to develop a robust and reliable estimation scheme to evaluate the in vivo lactate and lipid levels in glioma patients. The acquisition method combined J-difference modulation and reduced k-space sampling to obtain two set of 3D 1H-magnetic resonance spectroscopic imaging (MRSl) data, from which lactate editing was implemented. An iterative nonlinear least-squares algorithm for fitting Voigt line shape models in the frequency domain was developed for quantification for choline, creatine, NAA, lactate and lipid after non-parametric phase, frequency and baseline correction. Simulated data showed that the presented algorithm was robust and reliable. The quantification scheme was also applied to sixteen glioma patients prior to therapy.

Accurate diagnosis is critical for the treatment planning of brain tumors. At present, classification of tumor is based on histological examination of tissue samples. This is invasive and may be subject to sampling errors. Magnetic resonance spectroscopic imaging (MRSI) is a non-invasive technique that provides functional information and has been proposed as a tool for non-invasive tumor grading. The goal of this work is to find a classification method that (1) explicitly combines information from MRSI and MR imaging (MRI); (2) considers the MRSI characteristics of the entire lesion instead of a pre-selected region from within the anatomic lesion. Forty-nine newly-diagnosed glioma patients were studied with multivariate analysis based on recursive partitioning analysis (RPA) and linear discriminant analysis (LDA). The cross-validation classification error was 5 out of 49 patients. This suggested that characterizing the lesion by integrating the MRI/MRSI properties has the potential for improving the diagnosis and management of brain tumors.

A technique for displaying radiation treatment data with radiologic examinations within a picture archiving and communication system (PACS) framework is presented. It is shown to be technically feasible and of considerable use in diagnostic radiology.