A dictionary-based graph-cut algorithm for MRI reconstruction.

PURPOSE: Compressive sensing (CS)-based image reconstruction methods have proposed random undersampling schemes that produce incoherent, noise-like aliasing artifacts, which are easier to remove. The denoising process is critically assisted by imposing sparsity-enforcing priors. Sparsity is known to be induced if the prior is in the form of the Lp (0 ≤ p ≤ 1) norm. CS methods generally use a convex relaxation of these priors such as the L1 norm, which may not exploit the full power of CS. An efficient, discrete optimization formulation is proposed, which works not only on arbitrary Lp -norm priors as some non-convex CS methods do, but also on highly non-convex truncated penalty functions, resulting in a specific type of edge-preserving prior. These advanced features make the minimization problem highly non-convex, and thus call for more sophisticated minimization routines. THEORY AND METHODS: The work combines edge-preserving priors with random undersampling, and solves the resulting optimization using a set of discrete optimization methods called graph cuts. The resulting optimization problem is solved by applying graph cuts iteratively within a dictionary, defined here as an appropriately constructed set of vectors relevant to brain MRI data used here. RESULTS: Experimental results with in vivo data are presented. CONCLUSION: The proposed algorithm produces better results than regularized SENSE or standard CS for reconstruction of in vivo data.

Quantitative framework for prospective motion correction evaluation.

PURPOSE: Establishing a framework to evaluate performances of prospective motion correction (PMC) MRI considering motion variability between MRI scans. METHODS: A framework was developed to obtain quantitative comparisons between different motion correction setups, considering that varying intrinsic motion patterns between acquisitions can induce bias. Intrinsic motion was considered by replaying in a phantom experiment the recorded motion trajectories from subjects. T1-weighted MRI on five volunteers and two different marker fixations (mouth guard and nose bridge fixations) were used to test the framework. Two metrics were investigated to quantify the improvement of the image quality with PMC. RESULTS: Motion patterns vary between subjects as well as between repeated scans within a subject. This variability can be approximated by replaying the motion in a distinct phantom experiment and used as a covariate in models comparing motion corrections. We show that considering the intrinsic motion alters the statistical significance in comparing marker fixations. As an example, two marker fixations, a mouth guard and a nose bridge, were evaluated in terms of their effectiveness for PMC. A mouth guard achieved better PMC performance. CONCLUSION: Intrinsic motion patterns can bias comparisons between PMC configurations and must be considered for robust evaluations. A framework for evaluating intrinsic motion patterns in PMC is presented.

Imaging the microvessel caliber and density: Principles and applications of microvascular MRI.

Twenty years ago, theoretical developments were initiated to model the behavior of the NMR transverse relaxation rates in presence of vessels. These developments enabled the MRI-based mapping of mean vessel diameter, microvascular density, and vessel size index with comparable results to those obtained by a pathologist. The transfer of these techniques to routine clinical use has been hindered by the unavailability of the required sequences, namely fast gradient-echo spin-echo sequences. Based on the increasing accessibility of such sequences on MRI scanners over recent years, we review the principles governing microvascular MRI, the validation studies, and the applications that have been tested worldwide by several teams. We also provide some recommendations on how to measure microvessel caliber and density with MRI.

Numerical modeling of susceptibility-related MR signal dephasing with vessel size measurement: phantom validation at 3T.

PURPOSE: MRI is used to obtain quantitative oxygenation and blood volume information from the susceptibility-related MR signal dephasing induced by blood vessels. However, analytical models that fit the MR signal are usually not accurate over the range of small blood vessels. Moreover, recent studies have demonstrated limitations in the simultaneous assessment of oxygenation and blood volume. In this study, a multiparametric MRI framework that aims to measure vessel radii in addition to magnetic susceptibility and volume fraction was introduced. METHODS: The protocol consisted of gradient-echo sampling of the spin-echo, diffusion, T2, and B0 acquisitions. After correction steps, the data were postprocessed with a versatile numerical model of the MR signal. An important analytical model was implemented for comparison. The approach was validated in phantoms with coiling strings as proxy for blood vessels. RESULTS: The feasibility of the vessel radius measurement is demonstrated. The numerical model shows an improved accuracy compared with the analytical approach. However, both methods overestimate the radius. The simultaneous measurement of the magnetic susceptibility and the volume fraction remains challenging. CONCLUSION: The results suggest that this approach could be interesting in vivo to better characterize the microvasculature without contrast agent.

Is T2* enough to assess oxygenation? Quantitative blood oxygen level-dependent analysis in brain tumor.

PURPOSE: To analyze the contribution of the transverse relaxation parameter (T2), macroscopic field inhomogeneities (B0), and blood volume fraction (BVf) to blood oxygen level-dependent (BOLD)-based magnetic resonance (MR) measurements of blood oxygen saturation (SO2) obtained in a brain tumor model. MATERIALS AND METHODS: This study was approved by the local committee for animal care and use. Experiments were performed in accordance with permit 380 820 from the French Ministry of Agriculture. The 9L gliosarcoma cells were implanted in the brain of eight rats. Fifteen days later, 4.7-T MR examinations were performed to estimate T2*, T2, BVf, and T2*ΔB0corrected in the tumor and contralateral regions. MR estimates of SO2 were derived by combining T2, BVf, and T2*ΔB0corrected according to a recently described quantitative BOLD approach. Scatterplots and linear regression analysis were used to identify correlation between parameters. Paired Student t tests were used to compare the tumor region with the contralateral region. RESULTS: No significant correlations were found between T2* and any parameter in either tumor tissue or healthy tissue. T2* in the tumor and T2* in the uninvolved contralateral brain were the same (36 msec±4 [standard deviation] vs 36 msec±5, respectively), which might suggest similar oxygenation. Adding T2 information (98 msec±7 vs 68 msec±2, respectively) alone yields results that suggest apparent hypo-oxygenation of the tumor, while incorporating BVf (5.3%±0.6 vs 2.6%±0.3, respectively) alone yields results that suggest apparent hyperoxygenation. MR estimates of SO2 obtained with a complete quantitative BOLD analysis, although not correlated with T2* values, suggest normal oxygenation (68%±3 vs 65%±4, respectively). MR estimates of SO2 obtained in the contralateral tissue agree with previously reported values. CONCLUSION: Additional measurements, such as BVf, T2, and B0, are needed to obtain reliable information on oxygenation with BOLD MR imaging. The proposed quantitative BOLD approach, which includes these measurements, appears to be a promising tool with which to map tumor oxygenation.

Quantitative MR estimates of blood oxygenation based on T2*: a numerical study of the impact of model assumptions.

Several MR methods have been proposed over the last decade to obtain quantitative estimates of the tissue blood oxygen saturation (StO2) using a quantification of the blood oxygen level dependent effect. These approaches are all based on mathematical models describing the time evolution of the MR signal in biological tissues in the presence of magnetic field inhomogeneities. Although the experimental results are very encouraging, possible biases induced by the model assumptions have not been extensively studied. In this study, a numerical approach was used to examine the influence on T(2)*, blood volume fraction, and StO2 estimates of possible confounding factors such as water diffusion, intravascular signal, and presence of arterial blood in the voxel. To evaluate the impact of the vessel geometry, straight cylinders and realistic data from two-photon microscopy for microvascular geometry were compared. Our results indicate that the models are sufficiently realistic, based on a good correlation between ground truth and MR estimates of StO2.

Vessel size index measurements in a rat model of glioma: comparison of the dynamic (Gd) and steady-state (iron-oxide) susceptibility contrast MRI approaches.

Vessel size index (VSI), a parameter related to the distribution of vessel diameters, may be estimated using two MRI approaches: (i) dynamic susceptibility contrast (DSC) MRI following the injection of a bolus of Gd-chelate. This technique is routinely applied in the clinic to assess intracranial tissue perfusion in patients; (ii) steady-state susceptibility contrast with USPIO contrast agents, which is considered here as the standard method. Such agents are not available for human yet and the steady-state approach is currently limited to animal studies. The aim is to compare VSI estimates obtained with these two approaches on rats bearing C6 glioma (n = 7). In a first session, VSI was estimated from two consecutive injections of Gd-Chelate (Gd(1) and Gd(2)). In a second session (4 hours later), VSI was estimated using USPIO. Our findings indicate that both approaches yield comparable VSI estimates both in contralateral (VSI{USPIO} = 7.5 ± 2.0 µm, VSI{Gd(1)} = 6.5 ± 0.7 µm) and in brain tumour tissues (VSI{USPIO} = 19.4 ± 7.1 µm, VSI{Gd(1)} = 16.6 ± 4.5 µm). We also observed that, in the presence of BBB leakage (as it occurs typically in brain tumours), applying a preload of Gd-chelate improves the VSI estimate with the DSC approach both in contralateral (VSI{Gd(2)} = 7.1 ± 0.4 µm) and in brain tumour tissues (VSI{Gd(2)} = 18.5 ± 4.3 µm) but is not mandatory. VSI estimates do not appear to be sensitive to T(1) changes related to Gd extravasation. These results suggest that robust VSI estimates may be obtained in patients at 3 T or higher magnetic fields with the DSC approach.

Intracerebral injection of human mesenchymal stem cells impacts cerebral microvasculature after experimental stroke: MRI study.

Stroke, the leading cause of disability, lacks treatment beyond thrombolysis. The acute injection of human mesenchymal stem cells (hMSCs) provides a benefit which could be mediated by an enhancement of angiogenesis. A clinical autologous graft requires an hMSC culture delay incompatible with an acute administration. This study evaluates the cerebral microvascular changes after a delayed injection of hMSCs. At day 8 after middle cerebral artery occlusion (MCAo), two groups of rats received an intracerebral injection in the damaged brain of either 10 µL of cell suspension medium (MCAo-PBS, n = 4) or 4 × 105 hMSCs (MCAo-hMSC, n = 5). Two control groups of healthy rats underwent the same injection procedures in the right hemisphere (control-PBS, n = 6; control-hMSC, n = 5). The effect of hMSCs on the microvasculature was assessed by MRI using three parameters: apparent diffusion coefficient (ADC), cerebral blood volume (CBV) and vessel size index (VSI). At day 9, eight additional rats were euthanised for a histological study of the microvascular parameters (CBV, VSI and vascular fraction). No ADC difference was observed between MCAo groups. One day after intracerebral injection, hMSCs abolished the CBV increase observed in the lesion (MCAo-hMSC: 1.7 ± 0.1% versus MCAo-PBS: 2.2 ± 0.2%) and delayed the VSI increase (vasodilation) secondary to cerebral ischaemia. Histological analysis at day 9 confirmed that hMSCs modified the microvascular parameters (CBV, VSI and vascular fraction) in the lesion. No ADC, CBV or VSI differences were observed between control groups. At the stroke post-acute phase, hMSC intracerebral injection rapidly and transiently modifies the cerebral microvasculature. This microvascular effect can be monitored in vivo by MRI.

Knölker's iron complex: an efficient in situ generated catalyst for reductive amination of alkyl aldehydes and amines.

An aminated series: a well-defined iron-catalyzed reductive amination reaction of aldehydes and ketones with aliphatic amines using molecular hydrogen is presented. Under mild conditions, good yields for a broad range of alkyl ketones as well as aldehydes were achieved.

Evaluation of a quantitative blood oxygenation level-dependent (qBOLD) approach to map local blood oxygen saturation.

Blood oxygen saturation (SO(2)) is a promising parameter for the assessment of brain tissue viability in numerous pathologies. Quantitative blood oxygenation level-dependent (qBOLD)-like approaches allow the estimation of SO(2) by modelling the contribution of deoxyhaemoglobin to the MR signal decay. These methods require a high signal-to-noise ratio to obtain accurate maps through fitting procedures. In this article, we present a version of the qBOLD method at long TE taking into account separate estimates of T(2), total blood volume fraction (BV(f)) and magnetic field inhomogeneities. Our approach was applied to the brains of 13 healthy rats under normoxia, hyperoxia and hypoxia. MR estimates of local SO(2) (MR_LSO(2)) were compared with measurements obtained from blood gas analysis. A very good correlation (R(2) = 0.89) was found between brain MR_LSO(2) and sagittal sinus SO(2).

Impaired fMRI activation in patients with primary brain tumors.

To characterize peritumoral BOLD contrast disorders, 25 patients referred for resection of primary frontal or parietal neoplasms (low-grade glioma (LGG) (n=8); high-grade glioma (HGG) (n=7); meningioma (n=10)) without macroscopic tumoral infiltration of the primary sensorimotor cortex (SM1) were examined preoperatively using BOLD fMRI during simple motor tasks. Overall cerebral BOLD signal was estimated using vasoreactivity to carbogen inhalation. Using bolus of gadolinium, cerebral blood flow (CBF), cerebral blood volume (CBV), and mean transit time (MTT) were estimated. In a 1cm(3) region-of-interest centered on maximal T-value in SM1 contralateral to movements, interhemispheric asymmetry was evaluated using interhemispheric ratios for BOLD and perfusion parameters. During motor tasks contralateral to the tumor, ipsitumoral sensorimotor activations were decreased in HGG and meningiomas, correlated to the distance between the tumor and SM1. Whereas CBV was decreased in ipsitumoral SM1 for HGG, it remained normal in meningiomas. Changes in basal perfusion could not explain motor activation impairment in SM1. Decreased interhemispheric ratio of the BOLD response to carbogen was the best predictor to model the asymmetry of motor activation (R=0.51). Moreover, 94.9+/-4.9% of all motor activations overlapped significant BOLD response to carbogen inhalation.

Optimized 3D-NMR sampling for resonance assignment of partially unfolded proteins.

Resonance assignment of NMR spectra of unstructured proteins is made difficult by severe overlap due to the lack of secondary structure. Fortunately, this drawback is partially counterbalanced by the narrow line-widths due to the internal flexibility. Alternate sampling schemes can be used to achieve better resolution in less experimental time. Deterministic schemes (such as radial sampling) suffer however from the presence of systematic artifacts. Random acquisition patterns can alleviate this problem by randomizing the artifacts. We show in this communication that quantitative well-resolved spectra can be obtained, provided that the data points are properly weighted before FT. These weights can be evaluated using the concept of Voronoi cells associated with the data points. The introduced artifacts do not affect the direct surrounding of the peaks and thus do not alter the amplitude and frequency of the signals. This procedure is illustrated on 60-residue viral protein, which lacks any persistent secondary structure and thus exhibits major signal overlap.

Tissue oxygen saturation mapping with magnetic resonance imaging.

A quantitative estimate of cerebral blood oxygen saturation is of critical importance in the investigation of cerebrovascular disease. While positron emission tomography can map in vivo the oxygen level in blood, it has limited availability and requires ionizing radiation. Magnetic resonance imaging (MRI) offers an alternative through the blood oxygen level-dependent contrast. Here, we describe an in vivo and non-invasive approach to map brain tissue oxygen saturation (StO2) with high spatial resolution. StO2 obtained with MRI correlated well with results from blood gas analyses for various oxygen and hematocrit challenges. In a stroke model, the hypoxic areas delineated in vivo by MRI spatially matched those observed ex vivo by pimonidazole staining. In a model of diffuse traumatic brain injury, MRI was able to detect even a reduction in StO2 that was too small to be detected by histology. In a F98 glioma model, MRI was able to map oxygenation heterogeneity. Thus, the MRI technique may improve our understanding of the pathophysiology of several brain diseases involving impaired oxygenation.

A simulation tool for dynamic contrast enhanced MRI.

The quantification of bolus-tracking MRI techniques remains challenging. The acquisition usually relies on one contrast and the analysis on a simplified model of the various phenomena that arise within a voxel, leading to inaccurate perfusion estimates. To evaluate how simplifications in the interstitial model impact perfusion estimates, we propose a numerical tool to simulate the MR signal provided by a dynamic contrast enhanced (DCE) MRI experiment. Our model encompasses the intrinsic R1 and R2 relaxations, the magnetic field perturbations induced by susceptibility interfaces (vessels and cells), the diffusion of the water protons, the blood flow, the permeability of the vessel wall to the the contrast agent (CA) and the constrained diffusion of the CA within the voxel. The blood compartment is modeled as a uniform compartment. The different blocks of the simulation are validated and compared to classical models. The impact of the CA diffusivity on the permeability and blood volume estimates is evaluated. Simulations demonstrate that the CA diffusivity slightly impacts the permeability estimates (< 5% for classical blood flow and CA diffusion). The effect of long echo times is investigated. Simulations show that DCE-MRI performed with an echo time TE = 5 ms may already lead to significant underestimation of the blood volume (up to 30% lower for brain tumor permeability values). The potential and the versatility of the proposed implementation are evaluated by running the simulation with realistic vascular geometry obtained from two photons microscopy and with impermeable cells in the extravascular environment. In conclusion, the proposed simulation tool describes DCE-MRI experiments and may be used to evaluate and optimize acquisition and processing strategies.

Distribution and radiosensitizing effect of cholesterol-coupled Dbait molecule in rat model of glioblastoma.

BACKGROUND: Glioma is the most aggressive tumor of the brain and the most efficient treatments are based on radiotherapy. However, tumors are often resistant to radiotherapy due to an enhanced DNA repair activity. Short and stabilized DNA molecules (Dbait) have recently been proposed as an efficient strategy to inhibit DNA repair in tumor. METHODOLOGY/PRINCIPAL FINDINGS: The distribution of three formulations of Dbait, (i) Dbait alone, (ii) Dbait associated with polyethylenimine, and (iii) Dbait linked with cholesterol (coDbait), was evaluated one day after intratumoral delivery in an RG2 rat glioma model. Dbait molecule distribution was assessed in the whole organ with 2D-FRI and in brain sections. CoDbait was chosen for further studies given its good retention in the brain, cellular localization, and efficacy in inducing the activation of DNA repair effectors. The radiosensitizing effect of coDbait was studied in four groups of rats bearing RG2-glioma: no treatment, radiotherapy only, coDbait alone, and CoDbait with radiotherapy. Treatment started 7 days after tumor inoculation and consisted of two series of treatment in two weeks: coDbait injection followed by a selective 6-Gy irradiation of the head. We evaluated the radiosensitizing effect using animal survival, tumor volume, cell proliferation, and vasculature characteristics with multiparametric MRI. CoDbait with radiotherapy improved the survival of rats bearing RG2-glioma by reducing tumor growth and cell proliferation without altering tumor vasculature. CONCLUSION/SIGNIFICANCE: coDbait is therefore a promising molecular therapy to sensitize glioma to radiotherapy.

High-precision radiosurgical dose delivery by interlaced microbeam arrays of high-flux low-energy synchrotron X-rays.

Microbeam Radiation Therapy (MRT) is a preclinical form of radiosurgery dedicated to brain tumor treatment. It uses micrometer-wide synchrotron-generated X-ray beams on the basis of spatial beam fractionation. Due to the radioresistance of normal brain vasculature to MRT, a continuous blood supply can be maintained which would in part explain the surprising tolerance of normal tissues to very high radiation doses (hundreds of Gy). Based on this well described normal tissue sparing effect of microplanar beams, we developed a new irradiation geometry which allows the delivery of a high uniform dose deposition at a given brain target whereas surrounding normal tissues are irradiated by well tolerated parallel microbeams only. Normal rat brains were exposed to 4 focally interlaced arrays of 10 microplanar beams (52 microm wide, spaced 200 microm on-center, 50 to 350 keV in energy range), targeted from 4 different ports, with a peak entrance dose of 200Gy each, to deliver an homogenous dose to a target volume of 7 mm(3) in the caudate nucleus. Magnetic resonance imaging follow-up of rats showed a highly localized increase in blood vessel permeability, starting 1 week after irradiation. Contrast agent diffusion was confined to the target volume and was still observed 1 month after irradiation, along with histopathological changes, including damaged blood vessels. No changes in vessel permeability were detected in the normal brain tissue surrounding the target. The interlacing radiation-induced reduction of spontaneous seizures of epileptic rats illustrated the potential pre-clinical applications of this new irradiation geometry. Finally, Monte Carlo simulations performed on a human-sized head phantom suggested that synchrotron photons can be used for human radiosurgical applications. Our data show that interlaced microbeam irradiation allows a high homogeneous dose deposition in a brain target and leads to a confined tissue necrosis while sparing surrounding tissues. The use of synchrotron-generated X-rays enables delivery of high doses for destruction of small focal regions in human brains, with sharper dose fall-offs than those described in any other conventional radiation therapy.