TensorFit: A torch-based tool for ultrafast metabolite fitting of large MRSI data sets.

PURPOSE: To introduce a tool (TensorFit) for ultrafast and robust metabolite fitting of MRSI data based on Torch's auto-differentiation and optimization framework. METHODS: TensorFit was implemented in Python based on Torch's auto-differentiation to fit individual metabolites in MRS spectra. The underlying time domain and/or frequency domain fitting model is based on a linear combination of metabolite spectroscopic response. The computational time efficiency and accuracy of TensorFit were tested on simulated and in vivo MRS data and compared against TDFDFit and QUEST. RESULTS: TensorFit demonstrates a significant improvement in computation speed, achieving a 165-times acceleration compared with TDFDFit and 115 times against QUEST. TensorFit showed smaller percentual errors on simulated data compared with TDFDFit and QUEST. When tested on in vivo data, it performed similarly to TDFDFit with a 2% better fit in terms of mean squared error while obtaining a 169-fold speedup. CONCLUSION: TensorFit enables fast and robust metabolite fitting in large MRSI data sets compared with conventional metabolite fitting methods. This tool could boost the clinical applicability of large 3D MRSI by enabling the fitting of large MRSI data sets within computation times acceptable in a clinical environment.

Simultaneous multi-region detection of GABA+ and Glx using 3D spatially resolved SLOW-editing and EPSI-readout at 7T.

GABA+ and Glx (glutamate and glutamine) are widely studied metabolites, yet the commonly used magnetic resonance spectroscopy (MRS) techniques have significant limitations, including sensitivity to B0 and B1+-inhomogeneities, limited bandwidth of MEGA-pulses, high SAR which is accentuated at 7T. To address these limitations, we propose SLOW-EPSI method, employing a large 3D MRSI coverage and achieving a high resolution down to 0.26 ml. Simulation results demonstrate the robustness of SLOW-editing for both GABA+ and Glx against B0 and B1+-inhomogeneities within the range of [-0.3, +0.3] ppm and [40 %, 250 %], respectively. Two protocols, both utilizing a 70 mm thick FOV slab, were employed to target distinct brain regions in vivo, differentiated by their orientation: transverse and tilted. Protocol 1 (n = 11) encompassed 5 locations (cortical gray matter, white matter, frontal lobe, parietal lobe, and cingulate gyrus). Protocol 2 (n = 5) involved 9 locations (cortical gray matter, white matter, frontal lobe, occipital lobe, cingulate gyrus, caudate nucleus, hippocampus, putamen, and inferior thalamus). Quantitative analysis of GABA+ and Glx was conducted in a stepwise manner. First, B1+/B1--inhomogeneities were corrected using water reference data. Next, GABA+ and Glx values were calculated employing spectral fitting. Finally, the GABA+ level for each selected region was compared to the global Glx within the same subject, generating the GABA+/Glx_global ratio. Our findings from two protocols indicate that the GABA+/Glx_global level in cortical gray matter was approximately 16 % higher than in white matter. Elevated GABA+/Glx_global levels acquired with protocol 2 were observed in specific regions such as the caudate nucleus (0.118±0.067), putamen (0.108±0.023), thalamus (0.092±0.036), and occipital cortex (0.091±0.010), when compared to the cortical gray matter (0.079±0.012). Overall, our results highlight the effectiveness of SLOW-EPSI as a robust and efficient technique for accurate measurements of GABA+ and Glx at 7T. In contrast to previous SVS and 2D-MRSI based editing sequences with which only one or a limited number of brain regions can be measured simultaneously, the method presented here measures GABA+ and Glx from any brain area and any arbitrarily shaped volume that can be flexibly selected after the examination. Quantification of GABA+ and Glx across multiple brain regions through spectral fitting is achievable with a 9-minute acquisition. Additionally, acquisition times of 18-27 min (GABA+) and 9-18 min (Glx) are required to generate 3D maps, which are constructed using Gaussian fitting and peak integration.

Higher striatal glutamate in male youth with internet gaming disorder.

Internet gaming disorder (IGD) was included in the Diagnostic and Statistical Manual of Mental Disorders (DSM-5) as a research diagnosis, but little is known about its pathophysiology. Alterations in frontostriatal circuits appear to play a critical role in the development of addiction. Glutamate is considered an essential excitatory neurotransmitter in addictive disorders. This study's aim was to investigate striatal glutamate in youth with IGD compared to healthy controls (HC). Using a cross-sectional design, 25 adolescent male subjects fulfilling DSM-5 criteria for IGD and 26 HC, matched in age, education, handedness and smoking, were included in the analysis. A structural MPRAGE T1 sequence followed by a single-voxel magnetic resonance spectroscopy MEGA-PRESS sequence (TR = 1500 ms, TE = 68 ms, 208 averages) with a voxel size of 20 mm3 were recorded on 3 T Siemens Magnetom Prisma scanner. The voxel was placed in the left striatum. Group comparison of the relative glutamate and glutamine (Glx) was calculated using regression analysis. IGD subjects met an average of 6.5 of 9 DSM-5 IGD criteria and reported an average of 29 h of weekly gaming. Regression analysis showed a significant group effect for Glx, with higher Glx levels in IGD as compared to HC (coef. = .086, t (50) = 2.17, p = .035). Our study is the first to show higher levels of Glx in the striatum in youth with IGD. The elevation of Glx in the striatum may indicate hyperactivation of the reward system in IGD. Thus, results confirm that neurochemical alterations can be identified in early stages of behavioral addictions.

Deep learning pipeline for quality filtering of MRSI spectra.

With the rise of novel 3D magnetic resonance spectroscopy imaging (MRSI) acquisition protocols in clinical practice, which are capable of capturing a large number of spectra from a subject's brain, there is a need for an automated preprocessing pipeline that filters out bad-quality spectra and identifies contaminated but salvageable spectra prior to the metabolite quantification step. This work introduces such a pipeline based on an ensemble of deep-learning classifiers. The dataset consists of 36,338 spectra from one healthy subject and five brain tumor patients, acquired with an EPSI variant, which implemented a novel type of spectral editing named SLOtboom-Weng (SLOW) editing on a 7T MR scanner. The spectra were labeled manually by an expert into four classes of spectral quality as follows: (i) noise, (ii) spectra greatly influenced by lipid-related artifacts (deemed not to contain clinical information), (iii) spectra containing metabolic information slightly contaminated by lipid signals, and (iv) good-quality spectra. The AI model consists of three pairs of networks, each comprising a convolutional autoencoder and a multilayer perceptron network. In the classification step, the encoding half of the autoencoder is kept as a dimensionality reduction tool, while the fully connected layers are added to its output. Each of the three pairs of networks is trained on different representations of spectra (real, imaginary, or both), aiming at robust decision-making. The final class is assigned via a majority voting scheme. The F1 scores obtained on the test dataset for the four previously defined classes are 0.96, 0.93, 0.82, and 0.90, respectively. The arguably lower value of 0.82 was reached for the least represented class of spectra mildly influenced by lipids. Not only does the proposed model minimise the required user interaction, but it also greatly reduces the computation time at the metabolite quantification step (by selecting a subset of spectra worth quantifying) and enforces the display of only clinically relevant information.

α-D-Glucose as a non-radioactive MRS tracer for metabolic studies of the brain.

Changes in brain glucose metabolism occur in many neurological disorders as well as during aging. Most studies on the uptake of glucose in the brain use positron emission tomography, which requires injection of a radioactive tracer. Our study shows that ultra-high-field 1H-MRS can be used to measure α-D-glucose at 5.22 ppm in vivo, and the α-D-glucose can be used as a radiation-free tracer in the human brain.

Accurate prediction of isocitrate dehydrogenase -mutation status of gliomas using SLOW-editing magnetic resonance spectroscopic imaging at 7 T MR.

Background: 2-hydroxy-glutarate (2HG) is a metabolite that accumulates in isocitrate dehydrogenase (IDH)-mutated gliomas and can be detected noninvasively using MR spectroscopy. However, due to the low concentration of 2HG, established magnetic resonance spectroscopic imaging (MRSI) techniques at the low field have limitations with respect to signal-to-noise and to the spatial resolution that can be obtained within clinically acceptable measurement times. Recently a tailored editing method for 2HG detection at 7 Tesla (7 T) named SLOW-EPSI was developed. The underlying prospective study aimed to compare SLOW-EPSI to established techniques at 7 T and 3 T for IDH-mutation status determination. Methods: The applied sequences were MEGA-SVS and MEGA-CSI at both field strengths and SLOW-EPSI at 7 T only. Measurements were performed on a MAGNETOM-Terra 7 T MR-scanner in clinical mode using a Nova 1Tx32Rx head coil and on a 3 T MAGNETOM-Prisma scanner with a standard 32-channel head coil. Results: Fourteen patients with suspected glioma were enrolled. Histopathological confirmation was available in 12 patients. IDH mutation was confirmed in 9 out of 12 cases and 3 cases were characterized as IDH wildtype. SLOW-EPSI at 7 T showed the highest accuracy for IDH-status prediction (91.7% accuracy, 11 of the 12 predictions correct with 1 false negative case). At 7 T, MEGA-CSI had an accuracy of 58.3% and MEGA-SVS had an accuracy of 75%. At 3 T, MEGA-CSI showed an accuracy of 63.6% and MEGA-SVS of 33.3%. The co-edited cystathionine was detected in 2 out of 3 oligodendroglioma cases with 1p/19q codeletion. Conclusions: Depending on the pulse sequence, spectral editing can be a powerful tool for the noninvasive determination of the IDH status. SLOW-editing EPSI sequence is the preferable pulse sequence when used at 7 T for IDH-status characterization.

Submillimeter T1 atlas for subject-specific abnormality detection at 7T.

PURPOSE: Studies at 3T have shown that T1 relaxometry enables characterization of brain tissues at the single-subject level by comparing individual physical properties to a normative atlas. In this work, an atlas of normative T1 values at 7T is introduced with 0.6 mm isotropic resolution and its clinical potential is explored in comparison to 3T. METHODS: T1 maps were acquired in two separate healthy cohorts scanned at 3T and 7T. Using transfer learning, a template-based brain segmentation algorithm was adapted to ultra-high field imaging data. After segmenting brain tissues, volumes were normalized into a common space, and an atlas of normative T1 values was established by modeling the T1 inter-subject variability. A method for single-subject comparisons restricted to white matter and subcortical structures was developed by computing Z-scores. The comparison was applied to eight patients scanned at both field strengths for proof of concept. RESULTS: The proposed method for morphometry delivered segmentation masks without statistically significant differences from those derived with the original pipeline at 3T and achieved accurate segmentation at 7T. The established normative atlas allowed characterizing tissue alterations in single-subject comparisons at 7T, and showed greater anatomical details compared with 3T results. CONCLUSION: A high-resolution quantitative atlas with an adapted pipeline was introduced and validated. Several case studies on different clinical conditions showed the feasibility, potential and limitations of high-resolution single-subject comparisons based on quantitative MRI atlases. This method in conjunction with 7T higher resolution broadens the range of potential applications of quantitative MRI in clinical practice.

SLOW: A novel spectral editing method for whole-brain MRSI at ultra high magnetic field.

PURPOSE: At ultra-high field (UHF), B1+ -inhomogeneities and high specific absorption rate (SAR) of adiabatic slice-selective RF-pulses make spatial resolved spectral-editing extremely challenging with the conventional MEGA-approach. The purpose of the study was to develop a whole-brain resolved spectral-editing MRSI at UHF (UHF, B0 ≥ 7T) within clinical acceptable measurement-time and minimal chemical-shift-displacement-artifacts (CSDA) allowing for simultaneous GABA/Glx-, 2HG-, and PE-editing on a clinical approved 7T-scanner. METHODS: Slice-selective adiabatic refocusing RF-pulses (2π-SSAP) dominate the SAR to the patient in (semi)LASER based MEGA-editing sequences, causing large CSDA and long measurement times to fulfill SAR requirements, even using SAR-minimized GOIA-pulses. Therefore, a novel type of spectral-editing, called SLOW-editing, using two different pairs of phase-compensated chemical-shift selective adiabatic refocusing-pulses (2π-CSAP) with different refocusing bandwidths were investigated to overcome these problems. RESULTS: Compared to conventional echo-planar spectroscopic imaging (EPSI) and MEGA-editing, SLOW-editing shows robust refocusing and editing performance despite to B1+ -inhomogeneity, and robustness to B0 -inhomogeneities (0.2 ppm ≥ ΔB0  ≥ -0.2 ppm). The narrow bandwidth (∼0.6-0.8 kHz) CSAP reduces the SAR by 92%, RF peak power by 84%, in-excitation slab CSDA by 77%, and has no in-plane CSDA. Furthermore, the CSAP implicitly dephases water, lipid and all the other signals outside of range (≥ 4.6 ppm and ≤1.4 ppm), resulting in additional water and lipid suppression (factors ≥ 1000s) at zero SAR-cost, and no spectral aliasing artifacts. CONCLUSION: A new spectral-editing has been developed that is especially suitable for UHF, and was successfully applied for 2HG, GABA+, PE, and Glx-editing within 10 min clinical acceptable measurement time.

Preprocessing, analysis and quantification in single-voxel magnetic resonance spectroscopy: experts' consensus recommendations.

Once an MRS dataset has been acquired, several important steps must be taken to obtain the desired metabolite concentration measures. First, the data must be preprocessed to prepare them for analysis. Next, the intensity of the metabolite signal(s) of interest must be estimated. Finally, the measured metabolite signal intensities must be converted into scaled concentration units employing a quantitative reference signal to allow meaningful interpretation. In this paper, we review these three main steps in the post-acquisition workflow of a single-voxel MRS experiment (preprocessing, analysis and quantification) and provide recommendations for best practices at each step.

Contribution of macromolecules to brain 1 H MR spectra: Experts' consensus recommendations.

Proton MR spectra of the brain, especially those measured at short and intermediate echo times, contain signals from mobile macromolecules (MM). A description of the main MM is provided in this consensus paper. These broad peaks of MM underlie the narrower peaks of metabolites and often complicate their quantification but they also may have potential importance as biomarkers in specific diseases. Thus, separation of broad MM signals from low molecular weight metabolites enables accurate determination of metabolite concentrations and is of primary interest in many studies. Other studies attempt to understand the origin of the MM spectrum, to decompose it into individual spectral regions or peaks and to use the components of the MM spectrum as markers of various physiological or pathological conditions in biomedical research or clinical practice. The aim of this consensus paper is to provide an overview and some recommendations on how to handle the MM signals in different types of studies together with a list of open issues in the field, which are all summarized at the end of the paper.

No Effect of Anodal tDCS on Verbal Episodic Memory Performance and Neurotransmitter Levels in Young and Elderly Participants.

Healthy ageing is accompanied by cognitive decline that affects episodic memory processes in particular. Studies showed that anodal transcranial direct current stimulation (tDCS) to the left dorsolateral prefrontal cortex (DLPFC) may counteract this cognitive deterioration by increasing excitability and inducing neuroplasticity in the targeted cortical region. While stimulation gains are more consistent in initial low performers, relying solely on behavioural measures to predict treatment benefits does not suffice for a reliable implementation of this method as a therapeutic option. Hence, an exploration of the underlying neurophysiological mechanisms regarding the differential stimulation effect is warranted. Glutamatergic metabolites (Glx) and γ-aminobutyric acid (GABA) are involved in learning and memory processes and can be influenced with tDCS; wherefore, they present themselves as potential biomarkers for tDCS-induced behavioural gains, which are affiliated with neuroplasticity processes. In the present randomized, double-blind, sham-controlled, crossover study, 33 healthy young and 22 elderly participants received anodal tDCS to their left DLPFC during the encoding phase of a verbal episodic memory task. Using MEGA-PRESS edited magnetic resonance spectroscopy (MRS), Glx and GABA levels were measured in the left DLPFC before and after the stimulation period. Further, we tested whether baseline performance and neurotransmitter levels predicted subsequent gains. No beneficial group effects of tDCS emerged in either verbal retrieval performances or neurotransmitter concentrations. Moreover, baseline performance levels did not predict stimulation-induced cognitive gains, nor did Glx or GABA levels. Nevertheless, exploratory analyses suggested a predictive value of the Glx : GABA ratio, with lower ratios at baseline indicating greater tDCS-related gains in delayed recall performance. This highlights the importance of further studies investigating neurophysiological mechanisms underlying previously observed stimulation-induced cognitive benefits and their respective interindividual heterogeneity.

Terminology and concepts for the characterization of in vivo MR spectroscopy methods and MR spectra: Background and experts' consensus recommendations.

With a 40-year history of use for in vivo studies, the terminology used to describe the methodology and results of magnetic resonance spectroscopy (MRS) has grown substantially and is not consistent in many aspects. Given the platform offered by this special issue on advanced MRS methodology, the authors decided to describe many of the implicated terms, to pinpoint differences in their meanings and to suggest specific uses or definitions. This work covers terms used to describe all aspects of MRS, starting from the description of the MR signal and its theoretical basis to acquisition methods, processing and to quantification procedures, as well as terms involved in describing results, for example, those used with regard to aspects of quality, reproducibility or indications of error. The descriptions of the meanings of such terms emerge from the descriptions of the basic concepts involved in MRS methods and examinations. This paper also includes specific suggestions for future use of terms where multiple conventions have emerged or coexisted in the past.

Correction to: Impact of phenylalanine on cognitive, cerebral, and neurometabolic parameters in adult patients with phenylketonuria (the PICO study): a randomized, placebo-controlled, crossover, noninferiority trial.

An amendment to this paper has been published and can be accessed via the original article.

Impact of phenylalanine on cognitive, cerebral, and neurometabolic parameters in adult patients with phenylketonuria (the PICO study): a randomized, placebo-controlled, crossover, noninferiority trial.

BACKGROUND: The population of adult patients with early-treated phenylketonuria (PKU) following newborn screening is growing substantially. The ideal target range of blood phenylalanine (Phe) levels in adults outside pregnancy is a matter of debate. Therefore, prospective intervention studies are needed to evaluate the effects of an elevated Phe concentration on cognition and structural, functional, and neurometabolic parameters of the brain. METHODS: The PICO (Phenylalanine and Its Impact on Cognition) Study evaluates the effect of a 4-week Phe load on cognition and cerebral parameters in adults with early-treated PKU in a double-blind, randomized, placebo-controlled, crossover, noninferiority trial. PARTICIPANTS: Thirty adult patients with early-treated PKU and 30 healthy controls comparable to patients with regard to age, sex, and educational level will be recruited from the University Hospitals Bern and Zurich, Switzerland. Patients are eligible for the study if they are 18 years of age or older and had PKU diagnosed after a positive newborn screening and were treated with a Phe-restricted diet starting within the first 30 days of life. INTERVENTION: The cross-over intervention consists of 4-week oral Phe or placebo administration in patients with PKU. The study design mimics a Phe-restricted and a Phe-unrestricted diet using a double-blinded, placebo-controlled approach. OBJECTIVES: The primary objective of the PICO Study is to prospectively assess whether a temporarily elevated Phe level influences cognitive performance (working memory assessed with a n-back task) in adults with early-treated PKU. As a secondary objective, the PICO Study will elucidate the cerebral (fMRI, neural activation during a n-back task; rsfMRI, functional connectivity at rest; DTI, white matter integrity; and ASL, cerebral blood flow) and neurometabolic mechanisms (cerebral Phe level) that accompany changes in Phe concentration. Cognition, and structural and functional parameters of the brain of adult patients with early-treated PKU will be cross-sectionally compared to healthy controls. All assessments will take place at the University Hospital Bern, Switzerland. RANDOMIZATION: Central randomization will be used to assign participants to the different treatment arms with age, sex, and center serving as the stratification factors. Randomization lists will be generated by an independent statistician. Blinding: All trial personnel other than the statistician generating the randomization list and the personnel at the facility preparing the interventional product are blinded to the assigned treatment. DISCUSSION: Using a combination of neuropsychological and neuroimaging data, the PICO Study will considerably contribute to improve the currently insufficient level of evidence on how adult patients with early-treated PKU should be managed. TRIAL REGISTRATION: The study is registered at clinicaltrials.gov (NCT03788343) on the 27th of December 2018, at kofam.ch (SNCTP000003117) on the 17th of December 2018, and on the International Clinical Trials Registry Platform of the WHO.

Analysis of metabolic abnormalities in high-grade glioma using MRSI and convex NMF.

Clinical use of MRSI is limited by the level of experience required to properly translate MRSI examinations into relevant clinical information. To solve this, several methods have been proposed to automatically recognize a predefined set of reference metabolic patterns. Given the variety of metabolic patterns seen in glioma patients, the decision on the optimal number of patterns that need to be used to describe the data is not trivial. In this paper, we propose a novel framework to (1) separate healthy from abnormal metabolic patterns and (2) retrieve an optimal number of reference patterns describing the most important types of abnormality. Using 41 MRSI examinations (1.5 T, PRESS, TE 135 ms) from 22 glioma patients, four different patterns describing different types of abnormality were detected: edema, healthy without Glx, active tumor and necrosis. The identified patterns were then evaluated on 17 MRSI examinations from nine different glioma patients. The results were compared against BraTumIA, an automatic segmentation method trained to identify different tumor compartments on structural MRI data. Finally, the ability to predict future contrast enhancement using the proposed approach was also evaluated.

Iatrogenic Vessel Dissection in Endovascular Treatment of Acute Ischemic Stroke.

PURPOSE: Knowledge about the localization and outcome of iatrogenic dissection (ID) during endovascular treatment of acute ischemic stroke (AIS) is limited. We aimed to determine the frequency, clinical aspects and morphology of ID in endovascular AIS treatment and to identify predictors of this complication. METHODS: Digital subtraction angiography (DSA) of ID carried out during endovascular treatment between January 2000 and March 2012 have been re-evaluated. The ID localization and morphology were analyzed and related to the interventional techniques. Baseline clinical and radiological findings, treatment modality and outcome were compared with patients without ID. RESULTS: Out of 866 patients 18 (2%) suffered an ID (44% female, median age 64 years). Localization was extracranial in 15 (83%, 14 internal carotid artery and 1 vertebral artery) and intracranial in 3 (17%; 1 vertebrobasilar dissection and 2 in the anterior circulation). Of the IDs 5 (28%) resulted in a high-degree, 3 (17%) in a moderate, 5 (28%) in a mild and 5 (28%) in no stenosis and 8 IDs were stented in the acute phase. At 3 months 7 (42%) patients had a favorable outcome (modified Rankin score mRS ≤ 2) and 6 (33%) patients had died. Patients with ID had a different stroke etiology (p = 0.041), were more likely to be smokers (44% versus 19%, p = 0.015) and were more likely to be treated with mechanical thrombectomy (100% versus 60%, p < 0.001). Although two ID patients had relevant complications, the outcome did not differ between the groups. CONCLUSION: The occurrence of ID is a rare complication of endovascular AIS treatment associated with smoking and mechanical thrombectomy.

On the relation between MR spectroscopy features and the distance to MRI-visible solid tumor in GBM patients.

PURPOSE: To improve the detection of peritumoral changes in GBM patients by exploring the relation between MRSI information and the distance to the solid tumor volume (STV) defined using structural MRI (sMRI). METHODS: Twenty-three MRSI studies (PRESS, TE 135 ms) acquired from different patients with untreated GBM were used in this study. For each MRSI examination, the STV was identified by segmenting the corresponding sMRI images using BraTumIA, an automatic segmentation method. The relation between different metabolite ratios and the distance to STV was analyzed. A regression forest was trained to predict the distance from each voxel to STV based on 14 metabolite ratios. Then, the trained model was used to determine the expected distance to tumor (EDT) for each voxel of the MRSI test data. EDT maps were compared against sMRI segmentation. RESULTS: The features showing abnormal values at the longest distances to the tumor were: %NAA, Glx/NAA, Cho/NAA, and Cho/Cr. These four features were also the most important for the prediction of the distances to STV. Each EDT value was associated with a specific metabolic pattern, ranging from normal brain tissue to actively proliferating tumor and necrosis. Low EDT values were highly associated with malignant features such as elevated Cho/NAA and Cho/Cr. CONCLUSION: The proposed method enables the automatic detection of metabolic patterns associated with different distances to the STV border and may assist tumor delineation of infiltrative brain tumors such as GBM.

Quality of clinical brain tumor MR spectra judged by humans and machine learning tools.

PURPOSE: To investigate and compare human judgment and machine learning tools for quality assessment of clinical MR spectra of brain tumors. METHODS: A very large set of 2574 single voxel spectra with short and long echo time from the eTUMOUR and INTERPRET databases were used for this analysis. Original human quality ratings from these studies as well as new human guidelines were used to train different machine learning algorithms for automatic quality control (AQC) based on various feature extraction methods and classification tools. The performance was compared with variance in human judgment. RESULTS: AQC built using the RUSBoost classifier that combats imbalanced training data performed best. When furnished with a large range of spectral and derived features where the most crucial ones had been selected by the TreeBagger algorithm it showed better specificity (98%) in judging spectra from an independent test-set than previously published methods. Optimal performance was reached with a virtual three-class ranking system. CONCLUSION: Our results suggest that feature space should be relatively large for the case of MR tumor spectra and that three-class labels may be beneficial for AQC. The best AQC algorithm showed a performance in rejecting spectra that was comparable to that of a panel of human expert spectroscopists. Magn Reson Med 79:2500-2510, 2018. © 2017 International Society for Magnetic Resonance in Medicine.

Differentiating enhancing multiple sclerosis lesions, glioblastoma, and lymphoma with dynamic texture parameters analysis (DTPA): A feasibility study.

PURPOSE: MR-imaging hallmarks of glioblastoma (GB), cerebral lymphoma (CL), and demyelinating lesions are gadolinium (Gd) uptake due to blood-brain barrier disruption. Thus, initial diagnosis may be difficult based on conventional Gd-enhanced MRI alone. Here, the added value of a dynamic texture parameter analysis (DTPA) in the differentiation between these three entities is examined. DTPA is an in-house software tool that incorporates the analysis of quantitative texture parameters extracted from dynamic susceptibility contrast-enhanced (DSCE) images. METHODS: Twelve patients with multiple sclerosis (MS), 15 patients with GB, and five patients with CL were included. The image analysis method focuses on the DSCE image time series during bolus passage. Three time intervals were examined: inflow, outflow, and reperfusion time interval. Texture maps were computed. From the DSCE image series, mean, difference, standard deviation, and variance texture parameters were calculated and statistically analyzed and compared between the pathologies. RESULTS: The texture parameters of the original DSCE image series for mean, standard deviation, and variance showed the most significant differences (P-value between <0.00 and 0.05) between pathologies. Further, the texture parameters related to the standard deviation or variance (both associated with tissue heterogeneity) revealed the strongest discriminations between the pathologies. CONCLUSION: We conclude that dynamic perfusion texture parameters as assessed by the DTPA method allow discriminating MS, GB, and CL lesions during the first passage of contrast. DTPA used in combination with classification algorithms has the potential to find the most likely diagnosis given a postulated differential diagnosis.

Improving labeling efficiency in automatic quality control of MRSI data.

PURPOSE: To improve the efficiency of the labeling task in automatic quality control of MR spectroscopy imaging data. METHODS: 28'432 short and long echo time (TE) spectra (1.5 tesla; point resolved spectroscopy (PRESS); repetition time (TR)= 1,500 ms) from 18 different brain tumor patients were labeled by two experts as either accept or reject, depending on their quality. For each spectrum, 47 signal features were extracted. The data was then used to run several simulations and test an active learning approach using uncertainty sampling. The performance of the classifiers was evaluated as a function of the number of patients in the training set, number of spectra in the training set, and a parameter α used to control the level of classification uncertainty required for a new spectrum to be selected for labeling. RESULTS: The results showed that the proposed strategy allows reductions of up to 72.97% for short TE and 62.09% for long TE in the amount of data that needs to be labeled, without significant impact in classification accuracy. Further reductions are possible with significant but minimal impact in performance. CONCLUSION: Active learning using uncertainty sampling is an effective way to increase the labeling efficiency for training automatic quality control classifiers. Magn Reson Med 78:2399-2405, 2017. © 2017 International Society for Magnetic Resonance in Medicine.