Search

PET attenuation correction (AC) on systems lacking CT/transmission scanning, such as dedicated brain PET scanners and hybrid PET/MRI, is challenging. Direct AC in image-space, wherein PET images corrected for attenuation and scatter are synthesized from nonattenuation corrected PET (PET-nonAC) images in an end-to-end fashion using deep learning approaches (DLAC) is evaluated for various radiotracers used in molecular neuroimaging studies. One hundred eighty brain PET scans acquired using 18F-FDG, 18F-DOPA, 18F-Flortaucipir (targeting tau pathology), and 18F-Flutemetamol (targeting amyloid pathology) radiotracers (40 + 5, training/validation + external test, subjects for each radiotracer) were included. The PET data were reconstructed using CT-based AC (CTAC) to generate reference PET-CTAC and without AC to produce PET-nonAC images. A deep convolutional neural network was trained to generate PET attenuation corrected images (PET-DLAC) from PET-nonAC. The quantitative accuracy of this approach was investigated separately for each radiotracer considering the values obtained from PET-CTAC images as reference. A segmented AC map (PET-SegAC) containing soft-tissue and background air was also included in the evaluation. Quantitative analysis of PET images demonstrated superior performance of the DLAC approach compared to SegAC technique for all tracers. Despite the relatively low quantitative bias observed when using the DLAC approach, this approach appears vulnerable to outliers, resulting in noticeable local pseudo uptake and false cold regions. Direct AC in image-space using deep learning demonstrated quantitatively acceptable performance with less than 9% absolute SUV bias for the four different investigated neuroimaging radiotracers. However, this approach is vulnerable to outliers which result in large local quantitative bias.

OBJECTIVE: Quantitative PET/MR imaging is challenged by the accuracy of synthetic CT (sCT) generation from MR images. Deep learning-based algorithms have recently gained momentum for a number of medical image analysis applications. In this work, a novel sCT generation algorithm based on deep learning adversarial semantic structure (DL-AdvSS) is proposed for MRI-guided attenuation correction in brain PET/MRI.MATERIALS AND METHODS: The proposed DL-AdvSS algorithm exploits the ASS learning framework to constrain the synthetic CT generation process to comply with the extracted structural features from CT images. The proposed technique was evaluated through comparison to an atlas-based sCT generation method (Atlas), previously developed for MRI-only or PET/MRI-guided radiation planning. Moreover, the commercial segmentation-based approach (Segm) implemented on the Philips TF PET/MRI system was included in the evaluation. Clinical brain studies of 40 patients who underwent PET/CT and MR imaging were used for the evaluation of the proposed method under a two-fold cross validation scheme.RESULTS: The accuracy of cortical bone extraction and CT value estimation were investigated for the three different methods. Atlas and DL-AdvSS exhibited similar cortical bone extraction accuracy resulting in a Dice coefficient of 0.78 ± 0.07 and 0.77 ± 0.07, respectively. Likewise, DL-AdvSS and Atlas techniques performed similarly in terms of CT value estimation in the cortical bone region where a mean error (ME) of less than -11 HU was obtained. The Segm approach led to a ME of -1025 HU. Furthermore, the quantitative analysis of corresponding PET images using the three approaches assuming the CT-based attenuation corrected PET (PETCTAC) as reference demonstrated comparative performance of DL-AdvSS and Atlas techniques with a mean standardized uptake value (SUV) bias less than 4% in 63 brain regions. In addition, less that 2% SUV bias was observed in the cortical bone when using Atlas and DL-AdvSS approaches. However, Segm resulted in 14.7 ± 8.9% SUV underestimation in the cortical bone.CONCLUSION: The proposed DL-AdvSS approach demonstrated competitive performance with respect to the state-of-the-art atlas-based technique achieving clinically tolerable errors, thus outperforming the commercial segmentation approach used in the clinic.

Purpose Methods Magnetic resonance imaging (MRI)-guided radiation therapy (RT) treatment planning is limited by the fact that the electron density distribution required for dose calculation is not readily provided by MR imaging. We compare a selection of novel synthetic CT generation algorithms recently reported in the literature, including segmentation-based, atlas-based and machine learning techniques, using the same cohort of patients and quantitative evaluation metrics. Six MRI-guided synthetic CT generation algorithms were evaluated: one segmentation technique into a single tissue class (water-only), four atlas-based techniques, namely, median value of atlas images (ALMedian), atlas-based local weighted voting (ALWV), bone enhanced atlas-based local weighted voting (ALWV-Bone), iterative atlas-based local weighted voting (ALWV-Iter), and a machine learning technique using deep convolution neural network (DCNN). Results Conclusions Organ auto-contouring from MR images was evaluated for bladder, rectum, bones, and body boundary. Overall, DCNN exhibited higher segmentation accuracy resulting in Dice indices (DSC) of 0.93 +/- 0.17, 0.90 +/- 0.04, and 0.93 +/- 0.02 for bladder, rectum, and bones, respectively. On the other hand, ALMedian showed the lowest accuracy with DSC of 0.82 +/- 0.20, 0.81 +/- 0.08, and 0.88 +/- 0.04, respectively. DCNN reached the best performance in terms of accurate derivation of synthetic CT values within each organ, with a mean absolute error within the body contour of 32.7 +/- 7.9 HU, followed by the advanced atlas-based methods (ALWV: 40.5 +/- 8.2 HU, ALWV-Iter: 42.4 +/- 8.1 HU, ALWV-Bone: 44.0 +/- 8.9 HU). ALMedian led to the highest error (52.1 +/- 11.1 HU). Considering the dosimetric evaluation results, ALWV-Iter, ALWV, DCNN and ALWV-Bone led to similar mean dose estimation within each organ at risk and target volume with less than 1% dose discrepancy. However, the two-dimensional gamma analysis demonstrated higher pass rates for ALWV-Bone, DCNN, ALMedian and ALWV-Iter at 1%/1 mm criterion with 94.99 +/- 5.15%, 94.59 +/- 5.65%, 93.68 +/- 5.53% and 93.10 +/- 5.99% success, respectively, while ALWV and water-only resulted in 86.91 +/- 13.50% and 80.77 +/- 12.10%, respectively. Overall, machine learning and advanced atlas-based methods exhibited promising performance by achieving reliable organ segmentation and synthetic CT generation. DCNN appears to have slightly better performance by achieving accurate automated organ segmentation and relatively small dosimetric errors (followed closely by advanced atlas-based methods, which in some cases achieved similar performance). However, the DCNN approach showed higher vulnerability to anatomical variation, where a greater number of outliers was observed with this method. Considering the dosimetric results obtained from the evaluated methods, the challenge of electron density estimation from MR images can be resolved with a clinically tolerable error.

In recent years, considerable research has focused on the biological effect of endocrine-disrupting chemicals. Bisphenol A (BPA) has been implicated as an endocrine-disrupting chemical (EDC) due to its ability to mimic the action of endogenous estrogenic hormones. The aim of this study was to assess the effect of perinatal exposure to BPA on cerebral structural development and metabolism after birth. BPA (1mg/l) was administered in the drinking water of pregnant dams from day 6 of gestation until pup weaning. At postnatal day 20, in vivo metabolite concentrations in the rat pup hippocampus were measured using high field proton magnetic resonance spectroscopy. Further, brain was assessed histologically for growth, gross morphology, glial and neuronal development and extent of myelination. Localized proton magnetic resonance spectroscopy ((1)H MRS) showed in the BPA-exposed rat a significant increase in glutamate concentration in the hippocampus as well as in the Glu/Asp ratio. Interestingly these two metabolites are metabolically linked together in the malate-aspartate metabolic shuttle. Quantitative histological analysis revealed that the density of NeuN-positive neurons in the hippocampus was decreased in the BPA-treated offspring when compared to controls. Conversely, the density of GFAP-positive astrocytes in the cingulum was increased in BPA-treated offspring. In conclusion, exposure to low-dose BPA during gestation and lactation leads to significant changes in the Glu/Asp ratio in the hippocampus, which may reflect impaired mitochondrial function and also result in neuronal and glial developmental alterations.

Obstructive sleep apnea (OSA) is a condition characterized by upper airway muscle atonia with continued diaphragmatic efforts, resulting in repeated airway obstructions, periods of intermittent hypoxia, large thoracic pressure changes, and substantial shifts in arterial pressure with breathing cessation and resumption. The hypoxic exposure and hemodynamic changes likely induce the structural and functional deficits found in multiple brain areas, as shown by magnetic resonance imaging (MRI) procedures. Altered cerebral blood flow (CBF) may contribute to these localized deficits; thus, we examined regional CBF, using arterial spin labeling procedures, in 11 OSA (age, 49.1. ±. 12.2 years; 7 male) and 16 control subjects (42.3. ±. 10.2 years; 6 male) with a 3.0-Tesla MRI scanner. CBF maps were calculated, normalized to a common space, and regional CBF values across the brain quantified. Lowered CBF values emerged near multiple bilateral brain sites in OSA, including the corticospinal tracts, superior cerebellar peduncles, and pontocerebellar fibers. Lateralized, decreased CBF appeared near the left inferior cerebellar peduncles, left tapetum, left dorsal fornix/stria terminalis, right medial lemniscus, right red nucleus, right midbrain, and midline pons. Regional CBF values in OSA are significantly reduced in major sensory and motor fiber systems and motor regulatory sites, especially in structures mediating motor coordination; those reductions are often lateralized. The asymmetric CBF declines in motor regulatory areas may contribute to loss of coordination between upper airway and diaphragmatic musculature, and lead to further damage in the syndrome. © 2013 Elsevier Ireland Ltd.

INTRODUCTION: Eddy currents induced by switching of magnetic field gradients can lead to distortions in short echo-time spectroscopy or diffusion weighted imaging. In small bore magnets, such as human head-only systems, minimization of eddy current effects is more demanding because of the proximity of the gradient coil to conducting structures. METHODS: In the present study, the eddy current behavior achievable on a recently installed 7 tesla-68 cm bore head-only magnet was characterized. RESULTS: Residual effects after compensation were shown to be on the same order of magnitude as those measured on two whole body systems (3 and 4.7 T), while using two to three fold increased gradient slewrates.

Knowledge of T-1 relaxation times can be important for accurate relative and absolute quantification of brain metabolites, for sensitivity optimizations, for characterizing molecular dynamics, and for studying changes induced by various pathological conditions. H-1 T-1 relaxation times of a series of brain metabolites, including J-coupled ones, were determined using a progressive saturation (PS) technique that was validated with an adiabatic inversion-recovery (IR) method. The H-1 T-1 relaxation times of 16 functional groups of the neurochemical profile were measured at 14.1T and 9.4T. Overall, the T-1 relaxation times found at 14.1T were, within the experimental error, identical to those at 9.4T. The T(1)s of some coupled spin resonances of the neurochemical profile were measured for the first time (e.g., those of gamma-aminobutyrate [GABA], aspartate [Asp], alanine [Ala], phosphoethanolamine [PE], glutathione [GSH], N-acety-laspartylglutamate [NAAG], and glutamine [Gin]). Our results suggest that T-1 does not increase substantially beyond 9.4T. Furthermore, the similarity of T, among the metabolites (similar to 1.5 s) suggests that T-1 relaxation time corrections for metabolite quantification are likely to be similar when using rapid pulsing conditions. We therefore conclude that the putative T-1 increase of metabolites has a minimal impact on sensitivity when increasing B-0 beyond 9.4T. Magn Reson Med 62:862-867, 2009. (C) 2009 Wiley-Liss, Inc.

In in vivo (1)H spectroscopy, the signal at 1.32 ppm is usually assigned to lactate. This resonance position is shared with threonine at physiological pH. The similarity of spectral patterns of lactate and threonine renders the separate measurement of either threonine or lactate without and even with editing technically challenging. In this study, the threonine signal was detected using a single-shot multiple-bond editing technique and quantified in vivo in both rat and human brains. A threonine concentration was estimated at 0.8 +/- 0.3 mM (mean +/- SD, n = 6) in the rat brain and at approximately 0.33 mM in the human brain.

Knowledge of proton T-2 relaxation time of metabolites is essential for proper quantitation of metabolite concentrations in localized proton spectroscopy, especially at moderate to long TEs. Although the T-2 relaxation time of singlets, such as that of creatine and N-acetylaspartate, has been characterized in several studies, similar information is lacking from coupled spin resonances of cerebral metabolites. In this study, the T-2 relaxation time of coupled spin resonances and singlet resonances of cerebral metabolites was measured in rat brain in vivo at 9.4T. Spectra were acquired at 11 TEs using the SPin ECho, full Intensity Acquired Localized (SPECIAL) spectroscopy method. Data analysis was performed in the frequency domain with the LCModel software using simulated TE-specific basis sets. The T-2 relaxation times in compounds showing singlet resonances were 113 +/- 3 ms (total creatine), 178 +/- 29 ms (total choline) and 202 +/- 12 ms (N-acetylaspartate). The T-2 values of J-coupled metabolites ranged from 89 +/- 8 ms (glutamate) to 148 +/- 14 ms (myo-inositol). Copyright (C) 2007 John Wiley & Sons, Ltd.

Despite obvious improvements in spectral resolution at high magnetic field, the detection of C-13 labeling by H-1-[C-13] NMR spectroscopy remains hampered by spectral overlap, such as in the spectral region of H-1 resonances bound to C3 of glutamate (Glu) and glutamine (Gin), and C6 of N-acetylaspartate (NAA). The aim of this study was to develop, implement, and apply a novel H-1-[C-13] NMR spectroscopic editing scheme, dubbed "selective Resonance suppression by Adiabatic Carbon Editing and Decoupling single-voxel STimulated Echo Acquisition Mode" (RACED-STEAM). The sequence is based on the application of two asymmetric narrow-transition-band adiabatic RF inversion pulses at the resonance frequency of the C-13 coupled to the protons that need to be suppressed during the mixing time (TM) period, alternating the inversion band downfield and upfield from the C-13 resonance on odd and even scans, respectively, thus suppressing the detection of H-1 resonances bound to C-13 within the transition band of the inversion pulse. The results demonstrate the efficient suppression of H-1 resonances bound to C3 of Glu and Gin, and C4 of Glu, which allows the H-1 resonances bound to C6 of NAA and C4 of Gin to be revealed. The measured time course of the resolved labeling into NAA C6 with the new scheme was consistent with the slow turnover of NAA. Magn Reson Med 61:260-266, 2009. (c) 2009 Wiley-Liss, Inc.

Brain glutamine synthetase (GS) is an integral part of the glutamate-glutamine cycle and occurs in the glial compartment. In vivo Magnetic Resonance Spectroscopy (MRS) allows noninvasive measurements of the concentrations and synthesis rates of metabolites. N-15 MRS is an alternative approach to C-13 MRS. Incorporation of labeled N-15 from ammonia in cerebral glutamine allows to measure several metabolic reactions related to nitrogen metabolism, including the glutamate-glutamine cycle. To measure N-15 incorporation into the position 5N of glutamine and position 2N of glutamate and glutamine, we developed a novel N-15 pulse sequence to simultaneously detect, for the first time, [5-N-15] Gln and [2-N-15]Gln + Glu in vivo in the rat brain. In addition, we also measured for the first time in the same experiment localized H-1 spectra for a direct measurement of the net glutamine accumulation. Mathematical modeling of H-1 and N-15 MRS data allowed to reduce the number of assumptions and provided reliable determination of GS (0.30 +/- 0.050 mu mol/g per minute), apparent neurotransmission (0.26 +/- 0.030 mu mol/g per minute), glutamate dehydrogenase (0.029 +/- 0.002 mu mol/g per minute), and net glutamine accumulation (0.033 +/- 0.001 mu mol/g per minute). These results showed an increase of GS and net glutamine accumulation under hyperammonemia, supporting the concept of their implication in cerebral ammonia detoxification. Journal of Cerebral Blood Flow & Metabolism (2012) 32, 696-708; doi:10.1038/jcbfm.2011.173; published online 14 December 2011

Ultra-short echo-time proton single voxel spectra of rat brain were obtained on a 14.1 T 26 cm horizontal bore system. At this field, the fitted linewidth in the brain tissue of adult rats was about 11 Hz. New, separated resonances ascribed to phosphocholine, glycerophosphocholine and N-acetylaspartate were detected for the first time in vivo in the spectral range of 4.2-4.4 ppm. Moreover, improved separation of the resonances of lactate, alanine, gamma-aminobutyrate, glutamate and glutathione was observed. Metabolite concentrations were estimated by fitting in vivo spectra to a linear combination of simulated spectra of individual metabolites and a measured spectrum of macromolecules (LCModel). The calculated concentrations of metabolites were generally in excellent agreement with those obtained at 9.4 T. These initial results further indicated that increasing magnetic field strength to 14.1 T enhanced spectral resolution in H-1 NMR spectroscopy. This implies that the quantification of the neurochemical profile in rodent brain can be achieved with improved accuracy and precision. (C) 2008 Elsevier Inc. All rights reserved.

Full signal intensity H-1-[C-13] NMR spectroscopy, combining a preceding C-13-editing block based on an inversion BISEP (B-1-insensitive spectral editing pulse) with a spin-echo coherence-based localization, was developed and implemented at 14.1 T. C-13 editing of the proposed scheme was achieved by turning on and off the C-13 adiabatic full passage in the C-13-editing block to prepare inverted and noninverted C-13-coupled H-1 coherences along the longitudinal axis prior to localization. The novel H-1-[C-13] NMR approach was applied in vivo under infusion of the glia-specific substrate [2-C-13] acetate. Besides a similar to 50% improvement in sensitivity, spectral dispersion was enhanced at 14.1 T, especially for J-coupled metabolites such as glutamate and glutamine. A more distinct spectral structure at 1.9-2.2 ppm(parts per million) was observed, e.g., glutamate C3 showed a doublet pattern in both simulated H-1 spectrum and in vivo C-13-edited H-1 NMR spectra. Besides C-13 time courses of glutamate C4 and glutamine C4, the time courses of glutamate C3 and glutamine C3 obtained by H-1-[C-13] NMR spectroscopy were reported for the first time. Such capability should greatly improve the ability to study neuronglial metabolism using H-1-observed C-13-edited NMR spectroscopy. Magn Reson Med 64:334-340, 2010. (C) 2010 Wiley-Liss, Inc.

In this study, hypothalamic activation was performed by dehydration-induced anorexia (DIA) and overnight food suppression (OFS) in female rats. The assessment of the hypothalamic response to these challenges by manganese-enhanced MRI showed increased neuronal activity in the paraventricular nuclei (PVN) and lateral hypothalamus (LH), both known to be areas involved in the regulation of food intake. The effects of DIA and OFS were compared by generating T-score maps. Increased neuronal activation was detected in the PVN and LH of DIA rats relative to OFS rats. In addition, the neurochemical profile of the PVN and LH were measured by 1H MRS at 14.1T. Significant increases in metabolite levels were measured in DIA and OFS relative to control rats. Statistically significant increases in γ-aminobutyric acid were found in DIA (p=0.0007) and OFS (p

In the last decade, evidence has emerged indicating that the growth of a vast majority of tumors including gliomas is sustained by a subpopulation of cancer cells with stem cell properties called cancer initiating cells. These cells are able to initiate and propagate tumors and constitute only a fraction of all tumor cells. In the present study, we showed that intracerebral injection of cultured glioma-initiating cells into nude mice produced fast growing tumors showing necrosis and gadolinium enhancement in MR images, whereas gliomas produced by injecting freshly purified glioma-initiating cells grew slowly and showed no necrosis and very little gadolinium enhancement. Using proton localized spectroscopy at 14.1 Tesla, decreasing trends of N-acetylaspartate, glutamate and glucose concentrations and an increasing trend of glycine concentration were observed near the injection site after injecting cultured glioma-initiating cells. In contrast to the spectra of tumors grown from fresh cells, those from cultured cells showed intense peaks of lipids, increased absolute concentrations of glycine and choline-containing compounds, and decreased concentrations of glutamine, taurine and total creatine, when compared with a contralateral non-tumor-bearing brain tissue. A decrease in concentrations of N-acetylaspartate and -aminobutyrate was found in both tumor phenotypes after solid tumor formation. Further investigation is needed to determine the cause of the dissimilarities between the tumors grown from cultured glioma-initiating cells and those from freshly purified glioma-initiating cells, both derived from human glioblastomas. Copyright (C) 2011 John Wiley & Sons, Ltd.