Increased Water Content in Periventricular Caps in Patients without Acute Hydrocephalus.

BACKGROUND AND PURPOSE: Periventricular caps are a common finding on MR imaging and are believed to reflect focally increased interstitial water content due to dysfunctional transependymal transportation rather than ischemic-gliotic changes. We compared the quantitative water content of periventricular caps and microvascular white matter lesions, hypothesizing that periventricular caps associated with increased interstitial fluid content display higher water content than white matter lesions and are therefore differentiable from microvascular white matter lesions by measurement of the water content. MATERIALS AND METHODS: In a prospective study, we compared the water content of periventricular caps and white matter lesions in 50 patients using a quantitative multiple-echo, gradient-echo MR imaging water-mapping sequence. RESULTS: The water content of periventricular caps was significantly higher than that of white matter lesions (P = .002). Compared with normal white matter, the mean water content of periventricular caps was 17% ± 5% higher and the mean water content of white matter lesions was 11% ± 4% higher. Receiver operating characteristic analysis revealed that areas in which water content was 15% higher compared with normal white matter correspond to periventricular caps rather than white matter lesions, with a specificity of 93% and a sensitivity of 60% (P < .001). There was no significant correlation between the water content of periventricular caps and whole-brain volume (P = .275), white matter volume (P = .243), gray matter volume (P = .548), lateral ventricle volume (P = .800), white matter lesion volume (P = .081), periventricular cap volume (P = .081), and age (P = .224). CONCLUSIONS: Quantitative MR imaging allows differentiation between periventricular caps and white matter lesions. Water content quantification of T2-hyperintense lesions may be a useful additional tool for the characterization and differentiation of T2-hyperintense diseases.

Quality-based UnwRap of SUbdivided Large Arrays (URSULA) for high-resolution MRI data.

In Magnetic Resonance Imaging, mapping of the static magnetic field and the magnetic susceptibility is based on multidimensional phase measurements. Phase data are ambiguous and have to be unwrapped to their true range in order to exhibit a correct representation of underlying features. High-resolution imaging at ultra-high fields, where susceptibility and phase contrast are natural tools, can generate large datasets, which tend to dramatically increase computing time demands for spatial unwrapping algorithms. This article describes a novel method, URSULA, which introduces an artificial volume compartmentalisation that allows large-scale unwrapping problems to be broken down, making URSULA ideally suited for computational parallelisation. In the presented study, URSULA is illustrated with a quality-guided unwrapping approach. Validation is performed on numerical data and an application on a high-resolution measurement, at the clinical field strength of 3T is demonstrated. In conclusion, URSULA allows for a reduction of the problem size, a substantial speed-up and for handling large data sets without sacrificing the overall accuracy of the resulting phase information.

Hybrid MR-PET of brain tumours using amino acid PET and chemical exchange saturation transfer MRI.

PURPOSE: PET using radiolabelled amino acids has become a promising tool in the diagnostics of gliomas and brain metastasis. Current research is focused on the evaluation of amide proton transfer (APT) chemical exchange saturation transfer (CEST) MR imaging for brain tumour imaging. In this hybrid MR-PET study, brain tumours were compared using 3D data derived from APT-CEST MRI and amino acid PET using O-(2-18F-fluoroethyl)-L-tyrosine (18F-FET). METHODS: Eight patients with gliomas were investigated simultaneously with 18F-FET PET and APT-CEST MRI using a 3-T MR-BrainPET scanner. CEST imaging was based on a steady-state approach using a B1 average power of 1µT. B0 field inhomogeneities were corrected a Prametric images of magnetisation transfer ratio asymmetry (MTRasym) and differences to the extrapolated semi-solid magnetisation transfer reference method, APT# and nuclear Overhauser effect (NOE#), were calculated. Statistical analysis of the tumour-to-brain ratio of the CEST data was performed against PET data using the non-parametric Wilcoxon test. RESULTS: A tumour-to-brain ratio derived from APT# and 18F-FET presented no significant differences, and no correlation was found between APT# and 18F-FET PET data. The distance between local hot spot APT# and 18F-FET were different (average 20 ± 13 mm, range 4-45 mm). CONCLUSION: For the first time, CEST images were compared with 18F-FET in a simultaneous MR-PET measurement. Imaging findings derived from18F-FET PET and APT CEST MRI seem to provide different biological information. The validation of these imaging findings by histological confirmation is necessary, ideally using stereotactic biopsy.

Methods for molecular imaging of brain tumours in a hybrid MR-PET context: Water content, T2∗, diffusion indices and FET-PET.

The aim of this study is to present and evaluate a multiparametric and multi-modality imaging protocol applied to brain tumours and investigate correlations between these different imaging measures. In particular, we describe a method for rapid, non-invasive, quantitative imaging of water content of brain tissue, based on a single multiple-echo gradient-echo (mGRE) acquisition. We include in the processing a method for noise reduction of the multi-contrast data based on Principal Component Analysis (PCA). Noise reduction is a key ingredient to obtaining high-precision water content and transverse relaxation T2∗ values. The quantitative method is applied to brain tumour patients in a hybrid MR-PET environment. Active tumour tissue is identified by means of FET-PET; oedema, white and grey-matter segmentation is performed based on MRI contrasts. Water content information is not only relevant by itself, but also as a basis for correlations with other quantitative measures of water behaviour in tissue and interpreting the microenvironment of water. Water content in active tumour tissue (84%) and oedema (79%) regions is found to be higher than that of normal WM (69%) and close to that of normal GM (83%). Consistent with literature reports, mean kurtosis is measured to be lower in tumour and oedema regions than in normal WM and GM, whereas mean diffusivity is increased. Voxel-based correlations between water content and diffusion indices obtained with diffusion kurtosis tensor imaging, and between quantitative MRI and FET-PET are reported for 8 brain tumour patients. The effective transverse relaxation time T2∗ is found to be the MR parameter showing the strongest correlations with other MR indices derived here and with FET-PET.

MRI Appearance of Intracerebral Iodinated Contrast Agents: Is It Possible to Distinguish Extravasated Contrast Agent from Hemorrhage?

BACKGROUND AND PURPOSE: Hyperattenuated cerebral areas on postinterventional CT are a common finding after endovascular stroke treatment. There is uncertainty about the extent to which these hyperattenuated areas correspond to hemorrhage or contrast agent that extravasated into infarcted parenchyma during angiography. We evaluated whether it is possible to distinguish contrast extravasation from blood on MR imaging. MATERIALS AND METHODS: We examined the influence of iodinated contrast agents on T1, T2, and T2* and magnetic susceptibility in a phantom model and an ex vivo animal model. We determined T1, T2, and T2* relaxation times and magnetic susceptibility of iopamidol and iopromide in dilutions of 1:1; 1:2; 1:4; 1:10; and 1:100 with physiologic saline solution. We then examined the appearance of intracerebral iopamidol on MR imaging in an ex vivo animal model. To this end, we injected iopamidol into the brain of a deceased swine. RESULTS: Iopamidol and iopromide cause a negative susceptibility shift and T1, T2, and T2* shortening. The effects, however, become very small in dilutions of 1:10 and higher. Undiluted iopamidol, injected directly into the brain parenchyma, did not cause visually distinctive signal changes on T1-weighted spin-echo, T2-weighted turbo spin-echo, and T2*-weighted gradient recalled-echo imaging. CONCLUSIONS: It is unlikely that iodinated contrast agents extravasated into infarcted brain parenchyma cause signal changes that mimic hemorrhage on T1WI, T2WI, and T2*WI. Our results imply that extravasated contrast agents can be distinguished from hemorrhage on MR imaging.

Multimodal Imaging in Malignant Brain Tumors: Enhancing the Preoperative Risk Evaluation for Motor Deficits with a Combined Hybrid MRI-PET and Navigated Transcranial Magnetic Stimulation Approach.

BACKGROUND AND PURPOSE: Motor deficits in patients with brain tumors are caused mainly by irreversible infiltration of the motor network or by indirect mass effects; these deficits are potentially reversible on tumor removal. Here we used a novel multimodal imaging approach consisting of structural, functional, and metabolic neuroimaging to better distinguish these underlying causes in a preoperative setting and determine the predictive value of this approach. MATERIALS AND METHODS: Thirty patients with malignant brain tumors involving the central region underwent a hybrid O-(2-[(18)F]fluoroethyl)-L-tyrosine-PET-MR imaging and motor mapping by neuronavigated transcranial magnetic stimulation. The functional maps served as localizers for DTI tractography of the corticospinal tract. The spatial relationship between functional tissue (motor cortex and corticospinal tract) and lesion volumes as depicted by structural and metabolic imaging was analyzed. RESULTS: Motor impairment was found in nearly all patients in whom the contrast-enhanced T1WI or PET lesion overlapped functional tissue. All patients who functionally deteriorated after the operation showed such overlap on presurgical maps, while the absence of overlap predicted a favorable motor outcome. PET was superior to contrast-enhanced T1WI for revealing a motor deficit before the operation. However, the best correlation with clinical impairment was found for T2WI lesion overlap with functional tissue maps, but the prognostic value for motor recovery was not significant. CONCLUSIONS: Overlapping contrast-enhanced T1WI or PET-positive signals with motor functional tissue were highly indicative of motor impairment and predictive for surgery-associated functional outcome. Such a multimodal diagnostic approach may contribute to the risk evaluation of operation-associated motor deficits in patients with brain tumors.

Targeting murine heart and brain: visualisation conditions for multi-pinhole SPECT with (99m)Tc- and (123)I-labelled probes.

PURPOSE: The study serves to optimise conditions for multi-pinhole SPECT small animal imaging of (123)I- and (99m)Tc-labelled radiopharmaceuticals with different distributions in murine heart and brain and to investigate detection and dose range thresholds for verification of differences in tracer uptake. METHODS: A Triad 88/Trionix system with three 6-pinhole collimators was used for investigation of dose requirements for imaging of the dopamine D(2) receptor ligand [(123)I]IBZM and the cerebral perfusion tracer [(99m)Tc]HMPAO (1.2-0.4 MBq/g body weight) in healthy mice. The fatty acid [(123)I]IPPA (0.94 +/- 0.05 MBq/g body weight) and the perfusion tracer [(99m)Tc]sestamibi (3.8 +/- 0.45 MBq/g body weight) were applied to cardiomyopathic mice overexpressing the prostaglandin EP(3) receptor. RESULTS: In vivo imaging and in vitro data revealed 45 kBq total cerebral uptake and 201 kBq cardiac uptake as thresholds for visualisation of striatal [(123)I]IBZM and of cardiac [(99m)Tc]sestamibi using 100 and 150 s acquisition time, respectively. Alterations of maximal cerebral uptake of [(123)I]IBZM by >20% (116 kBq) were verified with the prerequisite of 50% striatal of total uptake. The labelling with [(99m)Tc]sestamibi revealed a 30% lower uptake in cardiomyopathic hearts compared to wild types. [(123)I]IPPA uptake could be visualised at activity doses of 0.8 MBq/g body weight. CONCLUSION: Multi-pinhole SPECT enables detection of alterations of the cerebral uptake of (123)I- and (99m)Tc-labelled tracers in an appropriate dose range in murine models targeting physiological processes in brain and heart. The thresholds of detection for differences in the tracer uptake determined under the conditions of our experiments well reflect distinctions in molar activity and uptake characteristics of the tracers.

Magnetic field dependence of the distribution of NMR relaxation times in the living human brain.

OBJECTIVE: This study investigates the field dependence of the distribution of in vivo, whole-brain T1 values, and its usefulness for white matter/grey matter segmentation. Results on T1 values are presented on 12 healthy volunteers. T2 and T2* distributions and their field dependence have been measured on the same cohort of volunteers. In this paper, however, only the T2 and T2* results on a single volunteer are presented. The reported field dependence of T2 and T2* values should, therefore, be given less weight than that of T1 times. MATERIALS AND METHODS: Relaxation times were measured in vivo on 12 healthy volunteers, using three nearly identical whole-body scanners, operating at field strengths of 1.5, 3, and 4 T and employing nearly identical software platforms and very similar hardware. T1 mapping was performed using TAPIR, a sequence based on the Look-Locker method. T2* mapping was performed with a multi-slice, multi-echo, gradient echo sequence. A multi-slice, multi-echo T2 mapping sequence based on the Carr-Purcell-Meiboom-Gill (CPMG) method was used to map T2. For each volunteer, the global distribution of T1 relaxation times was described as the superposition of three Gaussian distributions. The field and age-dependence of the centroids and widths of the three Gaussians was investigated. The segmentation of the brain in white and grey matter was performed separately for each field strength. Using the T1 segmentation and the fact that all maps were coregistered, we investigated the distribution of T2 and T*(2) values separately for the white and grey matter and described them with a Gaussian distribution in each case. RESULTS: Multi-slice quantitative maps were produced for the relaxation parameters T1 (near whole-brain coverage with 41 slices), T2* (whole-brain coverage, 55 slices), and T2 (27 slices). A clear age dependence was identified for grey matter T1 values and correlated with similar behaviour observed in a separate study of the brain water content. The increase with field strength of the bulk white and grey matter T1 values was well reproduced by both Bottomley's [1] and Fischer's [2] formulae, with parameters taken from the literature. The separation between the centroids was, however, either overestimated or underestimated by the two formulae. The width of the T1 distributions was found to increase with increasing field. CONCLUSIONS: The study of the field dependence of the NMR relaxation times is expected to allow for better differentiation between regions which are structurally different, provide a better insight into the microscopic structure of the brain and the molecular substrate of its function.