Common processing of in vivo MR spectra

Using Single-Voxel Magnetic Resonance Spectroscopy Data Acquired at 1.5T to Classify Multivoxel Data at 3T: A Proof-of-Concept Study

In vivo magnetic resonance spectroscopy (MRS) has two modalities, single-voxel (SV) and multivoxel (MV), in which one or more contiguous grids of SVs are acquired.

PURPOSE

To test whether MV grids can be classified with models trained with SV.

METHODS

Retrospective study. Training dataset: Multicenter multiformat SV INTERPRET, 1.5T. Testing dataset: MV eTumour, 3T. Two classification tasks were completed: 3-class (meningioma vs. aggressive vs. normal) and 4-class (meningioma vs. low-grade glioma vs. aggressive vs. normal). Five different methods were tested for feature selection. The classification was implemented using linear discriminant analysis (LDA), random forest, and support vector machines. The evaluation was completed with balanced error rate (BER) and area under the curve (AUC) on both sets. The accuracy in class prediction was calculated by developing a solid tumor index (STI) and segmentation accuracy with the Dice score.

RESULTS

The best method was sequential forward feature selection combined with LDA, with AUCs = 0.95 (meningioma), 0.89 (aggressive), 0.82 (low-grade glioma), and 0.82 (normal). STI was 66% (4-class task) and 71% (3-class task) because two cases failed completely and two more had suboptimal STI as defined by us.

DISCUSSION

The reasons for failure in the classification of the MV test set were related to the presence of artifacts.

Continuous Ingestion of Lacticaseibacillus rhamnosus JB-1 during Chronic Stress Ensures Neurometabolic and Behavioural Stability in Rats

The intestinal microbiome composition and dietary supplementation with psychobiotics can result in neurochemical alterations in the brain, which are possible due to the presence of the brain–gut–microbiome axis. In the present study, magnetic resonance spectroscopy (MRS) and behavioural testing were used to evaluate whether treatment with Lacticaseibacillus rhamnosus JB-1 (JB-1) bacteria alters brain metabolites’ levels and behaviour during continuous exposure to chronic stress. Twenty Wistar rats were subjected to eight weeks of a chronic unpredictable mild stress protocol. Simultaneously, half of them were fed with JB-1 bacteria, and the second half was given a daily placebo. Animals were examined at three-time points: before starting the stress protocol and after five and eight weeks of stress onset. In the elevated plus maze behavioural test the placebo group displayed increased anxiety expressed by almost complete avoidance of exploration, while the JB-1 dietary supplementation mitigated anxiety which resulted in a longer exploration time. Hippocampal MRS measurements demonstrated a significant decrease in glutamine + glutathione concentration in the placebo group compared to the JB-1 bacteria-supplemented group after five weeks of stress. With the progression of stress, the decrease of glutamate, glutathione, taurine, and macromolecular concentrations were observed in the placebo group as compared to baseline. The level of brain metabolites in the JB-1-supplemented rats were stable throughout the experiment, with only the taurine level decreasing between weeks five and eight of stress. These data indicated that the JB-1 bacteria diet might stabilize levels of stress-related neurometabolites in rat brain and could prevent the development of anxiety/depressive-like behaviour.

In vivo Human MR Spectroscopy Using a Clinical Scanner: Development, Applications, and Future Prospects

MR spectroscopy (MRS) is a unique and useful method for noninvasively evaluating biochemical metabolism in human organs and tissues, but its clinical dissemination has been slow and often limited to specialized institutions or hospitals with experts in MRS technology. The number of 3-T clinical MR scanners is now increasing, representing a major opportunity to promote the use of clinical MRS. In this review, we summarize the theoretical background and basic knowledge required to understand the results obtained with MRS and introduce the general consensus on the clinical utility of proton MRS in routine clinical practice. In addition, we present updates to the consensus guidelines on proton MRS published by the members of a working committee of the Japan Society of Magnetic Resonance in Medicine in 2013. Recent research into multinuclear MRS equipped in clinical MR scanners is explained with an eye toward future development. This article seeks to provide an overview of the current status of clinical MRS and to promote the understanding of when it can be useful. In the coming years, MRS-mediated biochemical evaluation is expected to become available for even routine clinical practice.

Comparison of two magnetic resonance imaging spectroscopy postprocessing methods

OBJECTIVE

The aim of this study was to compare the results obtained using SpectroView® (SV) and J-Magnetic Resonance User Interface (jMRUI) from the same magnetic resonance (MR) spectroscopy of hydrogen data.

METHODS

Data from 23 males with alcohol use disorder (AUD) and 23 healthy non-AUD males were acquired by a 1.5 Tesla MR using a PRESS sequence (TE=30 ms) in four voxels located in the right frontal and left frontal (RF and LF) lobes, and posterior cingulate (AC and PC). The ratio of the signals from both N-acetyl-aspartate (NAA) and choline (Cho) over creatine (Cr) was calculated automatically using SV and semiautomatically by an expert neuroradiologist using jMRUI. The software' agreement was calculated by the 95% limits of agreement (LoA) of the ratio of the obtained values.

RESULTS

The standard deviation was greater in jMRUI than in SV. Although there was a correlation between the results from both methods, it was not possible to predict their variance from one another. Additionally, the 95% LoA showed that jMRUI values were expected to vary from 38 to 190% of those obtained using SV for NAA/Cr in RF of AUD subjects and from 48 to 196% for NAA/Cr in CA of non-AUD individuals.

CONCLUSIONS

The difference between the methods may represent clinically significant magnitudes. We suggest the use of the same method when comparing spectroscopic data. We also suggest that in clinical practice, the automatic method should be preferred.

Nuclear magnetic resonance spectroscopy of human body fluids and in vivo magnetic resonance spectroscopy: Potential role in the diagnosis and management of prostate cancer

Prostate cancer is the most common solid organ cancer in men, and the second most common cause of male cancer-related mortality. It has few effective therapies, and is difficult to diagnose accurately. Prostate-specific antigen (PSA), which is currently the most effective diagnostic tool available, cannot reliably discriminate between different pathologies, and in fact only around 30% of patients found to have elevated levels of PSA are subsequently confirmed to actually have prostate cancer. As such, there is a desperate need for more reliable diagnostic tools that will allow the early detection of prostate cancer so that the appropriate interventions can be applied. Nuclear magnetic resonance (NMR) spectroscopy and magnetic resonance spectroscopy (MRS) are 2 high throughput, noninvasive analytical procedures that have the potential to enable differentiation of prostate cancer from other pathologies using metabolomics, by focusing specifically on certain metabolites which are associated with the development of prostate cancer cells and its progression. The value that this type of approach has for the early detection, diagnosis, prognosis, and personalized treatment of prostate cancer is becoming increasingly apparent. Recent years have seen many promising developments in the fields of NMR spectroscopy and MRS, with improvements having been made to hardware as well as to techniques associated with the acquisition, processing, and analysis of related data. This review focuses firstly on proton NMR spectroscopy of blood serum, urine, and expressed prostatic secretions in vitro, and then on 1- and 2-dimensional proton MRS of the prostate in vivo. Major advances in these fields and methodological principles of data collection, acquisition, processing, and analysis are described along with some discussion of related challenges, before prospects that proton MRS has for future improvements to the clinical management of prostate cancer are considered.

Processing tracking in jMRUI software for magnetic resonance spectra quantitation reproducibility assurance

Background

Proton magnetic resonance spectroscopy is a non-invasive measurement technique which provides information about concentrations of up to 20 metabolites participating in intracellular biochemical processes. In order to obtain any metabolic information from measured spectra a processing should be done in specialized software, like jMRUI. The processing is interactive and complex and often requires many trials before obtaining a correct result. This paper proposes a jMRUI enhancement for efficient and unambiguous history tracking and file identification.

Results

A database storing all processing steps, parameters and files used in processing was developed for jMRUI. The solution was developed in Java, authors used a SQL database for robust storage of parameters and SHA-256 hash code for unambiguous file identification. The developed system was integrated directly in jMRUI and it will be publically available. A graphical user interface was implemented in order to make the user experience more comfortable. The database operation is invisible from the point of view of the common user, all tracking operations are performed in the background.

Conclusions

The implemented jMRUI database is a tool that can significantly help the user to track the processing history performed on data in jMRUI. The created tool is oriented to be user-friendly, robust and easy to use. The database GUI allows the user to browse the whole processing history of a selected file and learn e.g. what processing lead to the results, where the original data are stored, to obtain the list of all processing actions performed on spectra.

From raw data to data-analysis for magnetic resonance spectroscopy--the missing link: jMRUI2XML

Background

Magnetic resonance spectroscopy provides metabolic information about living tissues in a non-invasive way. However, there are only few multi-centre clinical studies, mostly performed on a single scanner model or data format, as there is no flexible way of documenting and exchanging processed magnetic resonance spectroscopy data in digital format. This is because the DICOM standard for spectroscopy deals with unprocessed data.

This paper proposes a plugin tool developed for jMRUI, namely jMRUI2XML, to tackle the latter limitation. jMRUI is a software tool for magnetic resonance spectroscopy data processing that is widely used in the magnetic resonance spectroscopy community and has evolved into a plugin platform allowing for implementation of novel features.

Results

jMRUI2XML is a Java solution that facilitates common preprocessing of magnetic resonance spectroscopy data across multiple scanners. Its main characteristics are: 1) it automates magnetic resonance spectroscopy preprocessing, and 2) it can be a platform for outputting exchangeable magnetic resonance spectroscopy data. The plugin works with any kind of data that can be opened by jMRUI and outputs in extensible markup language format. Data processing templates can be generated and saved for later use.

The output format opens the way for easy data sharing- due to the documentation of the preprocessing parameters and the intrinsic anonymization - for example for performing pattern recognition analysis on multicentre/multi-manufacturer magnetic resonance spectroscopy data.

Conclusions

jMRUI2XML provides a self-contained and self-descriptive format accounting for the most relevant information needed for exchanging magnetic resonance spectroscopy data in digital form, as well as for automating its processing. This allows for tracking the procedures the data has undergone, which makes the proposed tool especially useful when performing pattern recognition analysis. Moreover, this work constitutes a first proposal for a minimum amount of information that should accompany any magnetic resonance processed spectrum, towards the goal of achieving better transferability of magnetic resonance spectroscopy studies.

Electronic supplementary material

The online version of this article (doi:10.1186/s12859-015-0796-5) contains supplementary material, which is available to authorized users.

Differences in the variability of cerebral proton magnetic resonance spectroscopy (1H-MRS) measurements within three HIV-infected cohorts

INTRODUCTION

Cerebral functional impairment remains prevalent in effectively treated HIV-infected subjects. As the results of formal cognitive testing are highly variable, surrogate markers to accurately measure cerebral function parameters are needed. Such markers include measurement of cerebral metabolite ratios (CMR) using proton magnetic resonance spectroscopy (1H-MRS). However, data on the inter-subject variability of CMR are sparse. Our aim was to assess inter-subject variability in CMRs within three different HIV-infected cohorts.

METHODS

Cerebral 1H-MRS was performed using a Phillips Achieva™ 1.5 Tesla magnetic resonance scanner in HIV-infected subjects as follows: 12 subjects before (group 1) and after intensification of antiretroviral therapy with maraviroc (group 2) and 13 subjects with acute viral hepatitis C (HCV) co-infection (group 3). The coefficients of variation (CV) for CMRs in each group were determined and compared using non-parametric tests to determine whether the inter-subject variability differed significantly. All baseline characteristics between the groups were similar.

RESULTS

Overall CVs for all CMRs in groups 1, 2 and 3 were 32.3%, 33.2% and 23.4%, respectively (group 1 vs. 2, p=0.863; group 1 vs. 3, p=0.076). On testing for differences in variability between individual CMRs, two metabolites in the right basal ganglia (RBG) had statistically significantly different CVs when comparing group 1 with group 3: N-acetyl aspartate/creatine (NAA/Cr), p=0.029 and myo-Inositol/creatine (mI/Cr), p=0.016.

CONCLUSION

The variability of 1H MRS-measurable CMRs in HIV-infected individuals was lower in those with acute HCV co-infection (group 3).We can conclude that the use of these CMRs in 1H MRS imaging in patients with HIV/acute HCV co-infection is more reliable to assess cerebral function than in patients with HIV infection alone. This has implications for future sample size estimations.

Metabolic profiling of dividing cells in live rodent brain by proton magnetic resonance spectroscopy (1HMRS) and LCModel analysis

Rationale

Dividing cells can be detected in the live brain by positron emission tomography or optical imaging. Here we apply proton magnetic resonance spectroscopy (¹HMRS) and a widely used spectral fitting algorithm to characterize the effect of increased neurogenesis after electroconvulsive shock in the live rodent brain via spectral signatures representing mobile lipids resonating at ∼1.30 ppm. In addition, we also apply the same ¹HMRS methodology to metabolically profile glioblastomas with actively dividing cells growing in RCAS-PDGF mice.

Methods

¹HMRS metabolic profiles were acquired on a 9.4T MRI instrument in combination with LCModel spectral analysis of: 1) rat brains before and after ECS or sham treatments and 2) RCAS-PDGF mice with glioblastomas and wild-type controls. Quantified ¹HMRS data were compared to post-mortem histology.

Results

Dividing cells in the rat hippocampus increased ∼3-fold after ECS compared to sham treatment. Quantification of hippocampal metabolites revealed significant decreases in N-acetyl-aspartate but no evidence of an elevated signal at ∼1.3 ppm (Lip13a+Lip13b) in the ECS compared to the sham group. In RCAS-PDGF mice a high density (22%) of dividing cells characterized glioblastomas. Nile Red staining revealed a small fraction (3%) of dying cells with intracellular lipid droplets in the tumors of RCAS-PDGF mice. Concentrations of NAA were lower, whereas lactate and Lip13a+Lip13b were found to be significantly higher in glioblastomas of RCAS-PDGF mice, when compared to normal brain tissue in the control mice.

Conclusions

Metabolic profiling using ¹HMRS in combination with LCModel analysis did not reveal correlation between Lip13a+Lip13b spectral signatures and an increase in neurogenesis in adult rat hippocampus after ECS. However, increases in Lip13a+Lip13b were evident in glioblastomas suggesting that a higher density of actively dividing cells and/or the presence of lipid droplets is necessary for LCModel to reveal mobile lipids.

Proton MR spectroscopy in metabolic assessment of musculoskeletal lesions

OBJECTIVE

The purposes of this review are to describe the principles and method of MR spectroscopy, summarize current published data on musculoskeletal lesions, and report additional cases that have been analyzed with recently developed quantitative methods.

CONCLUSION

Proton MR spectroscopy can be used to identify key tissue metabolites and may serve as a useful adjunct to radiographic evaluation of musculoskeletal lesions. A pooled analysis of 122 musculoskeletal tumors revealed that a discrete choline peak has a sensitivity of 88% and specificity of 68% in the detection of malignancy. Modest improvements in diagnostic accuracy in 22 of 122 cases when absolute choline quantification was used encourage the pursuit of development of choline quantification methods.

Effects of pyruvate dose on in vivo metabolism and quantification of hyperpolarized ¹³C spectra

Real-time in vivo measurements of metabolites are performed by signal enhancement of [1-(13)C]pyruvate using dynamic nuclear polarization, rapid dissolution and intravenous injection, acquisition of free induction decay signals and subsequent quantification of spectra. The commonly injected dose of hyperpolarized pyruvate is larger than typical tracer doses, with measurement before complete dilution of the injected bolus. Pyruvate is in exchange with its downstream metabolites lactate, alanine and bicarbonate. A transient exposure to high pyruvate blood concentrations may cause the saturation of cellular uptake and metabolic conversion. The goal of this study was to examine the effects of a [1-(13)C]pyruvate bolus on metabolic conversion in vivo. Spectra were quantified by three different methods: frequency-domain fitting with LCModel, time-domain fitting with AMARES and simple linear least-squares fitting in the time domain. Since the simple linear least-squares approach showed bleeding artifacts and LCModel produced noisier time signals. AMARES performed best in the quantification of in vivo hyperpolarized pyruvate spectra. We examined pyruvate doses of 0.1-0.4 mmol/kg (body mass) in male Wistar rats by acquiring slice-selective free induction decay signals in slices dominated by heart, liver and kidney tissue. Dose effects were noted in all cases, except for alanine in the cardiac slice below the dose of 0.2 mmol/kg. Our results indicate unlimited cellular uptake of pyruvate up to this dose and limited enzymatic activity of lactate dehydrogenase. In the cardiac slice above 0.2 mmol/kg and in liver and kidney slices, reflect limited cellular uptake or enzymatic activity, or a combination of both effects. The results indicate that the dose of pyruvate must be recognized as an important determinant for metabolic tissue kinetics, and saturation effects must be taken into account for the quantitative interpretation of the observed results.

Does acute hepatitis C infection affect the central nervous system in HIV-1 infected individuals?

Central nervous system (CNS) manifestations of chronic hepatitis C virus (HCV) and chronic human immune deficiency virus-1 (HIV-1) infections have been reported, but the impact of acute HCV infection on the CNS is unknown. A total of 10 individuals with chronic stable HIV-1 with documented acute HCV (HCV-RNA polymerase chain reaction positive and HCV antibody negative, group 1) underwent cerebral proton magnetic resonance spectroscopy (MRS) using acquisition parameters to quantify myo-inositol/creatine (mI/Cr) ratio in the right basal ganglia (RBG). Two matched control groups also underwent MRS; group 2: ten with chronic HIV-1 and no evidence of HCV, and group 3: ten with no evidence of HIV or HCV. Subjects also underwent computerized neurocognitive assessments (CogState). RBG mI/Cr ratio in group 1 (acute HCV in a background of HIV) was significantly lower than that in groups 2 and 3 [2.90 (+/-0.7) vs 3.34 (+/-0.4) and 3.43 (+/-0.4), mean (SD) for group 1 vs 2 and 3 respectively, P = 0.049], with 50% of subjects in group 1 having a mI/Cr ratio below the lowest observed ratio in either of the other groups. On neurocognitive testing, significant defects in the monitoring domain were observed in group-1, compared with matched controls (P = 0.021). Acute HCV in HIV-1 infected subjects is associated with CNS involvement. Clinicians should be vigilant of early CNS involvement when assessing subjects with acute HCV.

A sparse representation method for magnetic resonance spectroscopy quantification

In this paper, a sparse representation method is proposed for magnetic resonance spectroscopy (MRS) quantification. An observed MR spectrum is composed of a set of metabolic spectra of interest, a baseline and a noise. To separate the spectra of interest, the a priori knowledge about these spectra, such as signal models, the peak frequencies, and linewidth ranges of different resonances, is first integrated to construct a dictionary. The separation of the spectra of interest is then performed by using a pursuit algorithm to find their sparse representations with respect to the dictionary. For the challenging baseline problem, a wavelet filter is proposed to filter the smooth and broad components of both the observed spectra and the basis functions in the dictionary. The computation of sparse representation can then be carried out by using the remaining data. Simulation results show the good performance of this wavelet filtering-based strategy in separating the overlapping components between the baselines and the spectra of interest, when no appropriate model function for the baseline is available. Quantifications of in vivo brain MR spectra from tumor patients in different stages of progression demonstrate the effectiveness of the proposed method.

In vivo magnetic resonance spectroscopy: basic methodology and clinical applications

The clinical use of in vivo magnetic resonance spectroscopy (MRS) has been limited for a long time, mainly due to its low sensitivity. However, with the advent of clinical MR systems with higher magnetic field strengths such as 3 Tesla, the development of better coils, and the design of optimized radio-frequency pulses, sensitivity has been considerably improved. Therefore, in vivo MRS has become a technique that is routinely used more and more in the clinic. In this review, the basic methodology of in vivo MRS is described-mainly focused on (1)H MRS of the brain-with attention to hardware requirements, patient safety, acquisition methods, data post-processing, and quantification. Furthermore, examples of clinical applications of in vivo brain MRS in two interesting fields are described. First, together with a description of the major resonances present in brain MR spectra, several examples are presented of deviations from the normal spectral pattern associated with inborn errors of metabolism. Second, through examples of MR spectra of brain tumors, it is shown that MRS can play an important role in oncology.

IRIS-HSVD algorithm for automatic quantitation of in vivo 31P MRS

The rapid development of (31)P magnetic resonance spectroscopy (MRS) has enhanced non-invasive measurement of brain metabolites, which is important for biomedical research. The accuracy and efficiency of data post processing and quantification is paramount for MRS applications. One of the difficulties with in vivo(31)P MRS data quantification is the separation of broad line-width resonances from chemical compounds' resonances under a low signal-to-noise ratio condition. Furthermore, the chemical shift of some compounds caused by pH and Mg(2+) concentration can be troublesome. This work aims to develop an automatic algorithm using a state-space based quantification approach to solve the above mentioned problems. To achieve this aim, we utilized an HSVD based adaptive optimizing prior knowledge algorithm, which uses so called "interference" signals to optimize prior knowledge iteratively for parameter optimization. We termed this algorithm IRIS-HSVD, which stands for Iterative Reduction of Interference Signal HSVD. The Monte Carlo evaluations of the algorithm were conducted with simulated data using in vivo parameters commonly obtained from a 4T scanner. The performance of this algorithm using simulated data was compared to those of other automatic methods including HSVD and HTLS-PK. Examples of in vivo(31)P data obtained from brains of healthy subjects on a 4T MRI scanner were also presented, which demonstrated the superiority of the new method. The results were compared with those using AMARES.

MRS signal quantitation: a review of time- and frequency-domain methods

In this paper an overview of time-domain and frequency-domain quantitation methods is given. Advantages and drawbacks of these two families of quantitation methods are discussed. An overview of preprocessing methods, such as lineshape correction methods or unwanted component removal methods, is also given. The choice of the quantitation method depends on the data under investigation and the pursued objectives.

Spectrum separation resolves partial-volume effect of MRSI as demonstrated on brain tumor scans

Magnetic resonance spectroscopic imaging (MRSI) is currently used clinically in conjunction with anatomical MRI to assess the presence and extent of brain tumors and to evaluate treatment response. Unfortunately, the clinical utility of MRSI is limited by significant variability of in vivo spectra. Spectral profiles show increased variability because of partial coverage of large voxel volumes, infiltration of normal brain tissue by tumors, innate tumor heterogeneity, and measurement noise. We address these problems directly by quantifying the abundance (i.e. volume fraction) within a voxel for each tissue type instead of the conventional estimation of metabolite concentrations from spectral resonance peaks. This 'spectrum separation' method uses the non-negative matrix factorization algorithm, which simultaneously decomposes the observed spectra of multiple voxels into abundance distributions and constituent spectra. The accuracy of the estimated abundances is validated on phantom data. The presented results on 20 clinical cases of brain tumor show reduced cross-subject variability. This is reflected in improved discrimination between high-grade and low-grade gliomas, which demonstrates the physiological relevance of the extracted spectra. These results show that the proposed spectral analysis method can improve the effectiveness of MRSI as a diagnostic tool.

Cerebral immune activation in chronic hepatitis C infection: a magnetic resonance spectroscopy study

BACKGROUND/AIMS

Abnormal cerebral metabolism and cognitive impairments have been reported in patients with chronic hepatitis C (HCV) but studies have failed to demonstrate a relationship between these findings.

METHODS

Twenty-five HCV-positive patients with histologically-mild liver disease were studied with cerebral proton magnetic resonance spectroscopy (MRS), using acquisition parameters to quantify myo-inositol (mI) and other metabolites in frontal white matter (FWM). Patients underwent automated attention and working memory tests (Cognitive Drug Research test system).

RESULTS

The mean mI/ creatine ratio in the HCV+ve patients (0.64, SD 0.21) was significantly higher (p=0.02) than in healthy controls (0.52, SD 0.10). On cognitive testing, the HCV+ve patients showed impairments in 2/4 composite scores, reflecting working memory and attention, compared to normative data from healthy volunteers (p<0.005) and HCV-ve controls (p=0.03). There was a significant association between elevated FWM mI/creatine and prolonged working memory reaction times (R=0.72, p=0.002).

CONCLUSIONS

Elevated FWM mI/ creatine is a feature of HIV-related minor cognitive-motor disorder. It is associated with infection and immune activation of microglial cells. The similar findings in this study suggest that cerebral immune activation may also occur in HCV infection. This may underlie the mild neurocognitive impairment and neuropsychological symptoms observed in a proportion of patients.

Recent advances in magnetic resonance neurospectroscopy

Over the past two decades, proton magnetic resonance spectroscopy (proton MRS) of the brain has made the transition from research tool to a clinically useful modality. In this review, we first describe the localization methods currently used in MRS studies of the brain and discuss the technical and practical factors that determine the applicability of the methods to particular clinical studies. We also describe each of the resonances detected by localized solvent-suppressed proton MRS of the brain and discuss the metabolic and biochemical information that can be derived from an analysis of their concentrations. We discuss spectral quantitation and summarize the reproducibility of both single-voxel and multivoxel methods at 1.5 and 3-4 T. We have selected three clinical neurologic applications in which there has been a consensus as to the diagnostic value of MRS and summarize the information relevant to clinical applications. Finally, we speculate about some of the potential technical developments, either in progress or in the future, that may lead to improvements in the performance of proton MRS.

Chemical shift imaging of molecular transport in colloidal systems: Visualization and quantification of diffusion processes

Magnetic resonance imaging with chemical shift resolution is demonstrated to provide detailed information about molecular transport on the macroscopic scale in complex colloidal systems. The concentrations of species with distinct 1H resonance lines can be quantified from spatially resolved, high-resolution, 1H nuclear magnetic resonance spectra. The method is demonstrated by experiments on three systems with multiple simultaneous transport processes where the diffusion coefficients depend on position and/or on the concentration of other species: (1) release of poly(ethylene glycol) and imidazole from a hydrogel into an external reservoir of water, (2) migration of acetic acid and tetramethylammonium ions in a highly concentrated water-in-oil emulsion with initially non-uniform concentration of solutes, and (3) release of tetramethylammonium ions loaded into a hydrogel triggered by the diffusion of methyl green into the gel matrix.

1H MR spectroscopy of the brain: absolute quantification of metabolites

Hydrogen 1 (1H) magnetic resonance (MR) spectroscopy enables noninvasive in vivo quantification of metabolite concentrations in the brain. Currently, metabolite concentrations are most often presented as ratios (eg, relative to creatine) rather than as absolute concentrations. Despite the success of this approach, it has recently been suggested that relative quantification may introduce substantial errors and can lead to misinterpretation of spectral data and to erroneous metabolite values. The present review discusses relevant methods to obtain absolute metabolite concentrations with a clinical MR system by using single-voxel spectroscopy or chemical shift imaging. Important methodological aspects in an absolute quantification strategy are addressed, including radiofrequency coil properties, calibration procedures, spectral fitting methods, cerebrospinal fluid content correction, macromolecule suppression, and spectral editing. Techniques to obtain absolute concentrations are now available and can be successfully applied in clinical practice. Although the present review is focused on 1H MR spectroscopy of the brain, a large part of the methodology described can be applied to other tissues as well.

Optimal classification of long echo time in vivo magnetic resonance spectra in the detection of recurrent brain tumors

We describe the optimal high-level postprocessing of single-voxel (1)H magnetic resonance spectra and assess the benefits and limitations of automated methods as diagnostic aids in the detection of recurrent brain tumor. In a previous clinical study, 90 long-echo-time single-voxel spectra were obtained from 52 patients and classified during follow-up (30/28/32 normal/non-progressive tumor/tumor). Based on these data, a large number of evaluation strategies, including both standard resonance line quantification and algorithms from pattern recognition and machine learning, were compared in a quantitative evaluation. Results from linear and non-linear feature extraction, including ICA, PCA and wavelet transformations, and also the data from resonance line quantification were combined systematically with different classifiers such as LDA, chemometric methods (PLS, PCR), support vector machines and ensemble methods. Classification accuracy was assessed using a leave-one-out cross-validation scheme and the area under the curve (AUC) of the receiver operator characteristic (ROC). A regularized linear regression on spectra with binned channels reached 91% classification accuracy compared with 83% from quantification. Interpreting the loadings of these regressions, we find that lipid and lactate signals are too unreliable to be used in a simple machine rule. Choline and NAA are the main source of relevant information. Overall, we find that fully automated pattern recognition algorithms perform as well as, or slightly better than, a manually controlled and optimized resonance line quantification.

Quantifying in vivo MR spectra with circles

Accurate and robust quantification of in vivo magnetic resonance spectroscopy (MRS) data is essential to its application in research and medicine. The performance of existing analysis methods is problematic for in vivo studies where low signal-to-noise ratio, overlapping peaks and intense artefacts are endemic. Here, a new frequency-domain technique for MRS data analysis is introduced wherein the circular trajectories which result when spectral peaks are projected onto the complex plane, are fitted with active circle models. The use of active contour strategies naturally allows incorporation of prior knowledge as constraint energy terms. The problem of phasing spectra is eliminated, and baseline artefacts are dealt with using active contours-snakes. The stability and accuracy of the new technique, CFIT, is compared with a standard time-domain fitting tool, using simulated 31P data with varying amounts of noise and 98 real human chest and heart 31P MRS data sets. The real data were also analyzed by our standard frequency-domain absorption-mode technique. On the real data, CFIT demonstrated the least fitting failures of all methods and an accuracy similar to the latter method, with both these techniques outperforming the time-domain approach. Contrasting results from simulations argue that performance relative to Cramer-Rao Bounds may not be a suitable indicator of fitting performance with typical in vivo data such as these. We conclude that CFIT is a stable, accurate alternative to the best existing methods of fitting in vivo data.

Variations due to analysis technique in intracellular pH measurements in simulated and in vivo31P MR spectra of the human brain

PURPOSE

To investigate variation in pH generated by different analysis techniques and to find the most robust method, 31P MR brain spectra were acquired in vivo. Three different methods were used to measure the chemical shift of inorganic phosphate (Pi) relative to phosphocreatine (PCr).

MATERIALS AND METHODS

Eight healthy volunteers were scanned four times, and manual measurement of the chemical shift in a frequency domain spectrum using the manufacturer's software was compared with values produced by a frequency-domain analysis method (NMR1) and a prior-knowledge-based time-domain technique (MRUI). To explain the in vivo data, simulations of brain spectra, modified in ways typical of real variations in vivo, were produced and the pH was measured using manual measurement and MRUI.

RESULTS

Different measurement techniques produced systematically different pH values, with manual measurement producing the lowest variability (manual measurement: pH = 6.999, CoV = 0.297; NMR1: pH = 7.042, CoV = 0.501; MRUI: pH = 7.036, CoV = 0.606). While MRUI more accurately measured the pH of unaltered simulations, it was systematically affected by altering the simulated spectra. Manual measurement was unaffected.

CONCLUSION

Manual measurement produces the most consistent pH value, and there is no benefit in using more complex automated spectral fitting methods to measure the pH.

The fast Padé transform in magnetic resonance spectroscopy for potential improvements in early cancer diagnostics

The convergence rates of the fast Padé transform (FPT) and the fast Fourier transform (FFT) are compared. These two estimators are used to process a time-signal encoded at 4 T by means of one-dimensional magnetic resonance spectroscopy (MRS) for healthy human brain. It is found systematically that at any level of truncation of the full signal length, the clinically relevant resonances that determine concentrations of metabolites in the investigated tissue are significantly better resolved in the FPT than in the FFT. In particular, the FPT has a better resolution than the FFT for the same signal length. Moreover, the FPT can achieve the same resolution as the FFT by using twice shorter signals. Implications of these findings for two-dimensional magnetic resonance spectroscopy as well as for two- and three-dimensional magnetic resonance spectroscopic imaging are highlighted. Self-contained cross-validation of all the results from the FPT is secured by using two conceptually different, equivalent algorithms (inside and outside the unit-circle), that are both valid in the entire complex frequency plane. The difference between the results from these two variants of the FPT is indistinguishable from the background noise. This constitutes robust error analysis of proven validity. The FPT shows promise in applications of MRS for early cancer detection.

Prior knowledge for time domain quantification of in vivo brain or liver 31P MR spectra

Prior knowledge is required when quantifying in vivo (31)P magnetic resonance spectra from the brain or liver. The prior knowledge system we have used models both the phosphomonoester and phosphodiester resonances as two peaks of equal linewidth and fixed relative chemical shift. The analysis of the data is carried out in the time domain, which allows the broad component of the spectra to be modelled. This prior knowledge method has been tested for analysis of in vivo (31)P MR spectra from the liver and brain and gives results consistent with other methods that are also used to analyse the spectra, but with reduced variability. This technique may be utilized for studies requiring serial MR spectroscopy examinations, before and after patient treatment.

Automated correction of unwanted phase jumps in reference signals which corrupt MRSI spectra after eddy current correction

A commonly applied step in the postprocessing of gradient localized proton MR spectroscopy, is correction for eddy current effects using the water signal as a reference. However, this method can degrade some of the metabolite signals, in particular if applied on proton MR spectroscopic imaging data. This artifact arises from the water reference signal in the presence of a second signal which resonates close to the main water resonance. The interference of both resonances will introduce jumps in the phase of the reference time domain signal. Using this phase for eddy current correction will result in a ringing artifact in the frequency domain of the metabolite signal over the whole frequency range. We propose a moving window correction algorithm, which screens the phase of reference signals and removes phase jumps in time domain caused by interference of signals from multiple spin systems. The phase jumps may be abrupt or gradually distributed over several time data points. Because the correction algorithm only corrects time data points which contain phase jumps, the phase is minimally disrupted. Furthermore, the algorithm is automated for large datasets, correcting only those water reference signals which are corrupted. After correction of the corrupted reference signals, normal eddy current correction may be performed. The algorithm is compared with a method which uses a low-pass filter and tested on simulated data as well as on in vivo proton spectroscopic imaging data from a healthy volunteer and from patients with a brain tumor.