Anke Henning - Academia.edu (original) (raw)

Papers by Anke Henning

Magnetic Resonance in Medicine, 2020

This is an open access article under the terms of the Creative Commons Attribution License, which... more This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Heart, 2009

Background Patients with heart failure with preserved ejection fraction (HFpEF) have dyspnoea on ... more Background Patients with heart failure with preserved ejection fraction (HFpEF) have dyspnoea on exertion and limited exercise capacity. In this study we evaluated the role of exercise-related changes in left ventricular relaxation and of vasculoventricular coupling (VVC) as the mechanism of this limitation and whether cardiac energetic impairment may underlie these abnormalities. Methods We prospectively studied 37 patients with HFpEF. All had signs and/or symptoms of heart failure and a normal ejection fraction (>50%). Twenty age and gender-matched healthy volunteers were also studied. VVC and left ventricular filling characteristics were assessed at rest and on exercise by multiple uptake gated acquisition scan, and time to peak filling (nTTPF) was corrected for the RR interval (an indirect measure of the rate of left ventricular active relaxation) was derived. In vivo myocardial energetic state was assessed by 31P magnetic resonance spectroscopy at 3 Tesla. All subjects also ...

Introduction In vivo proton spectroscopy is a valuable tool that provides additional knowledge ab... more Introduction In vivo proton spectroscopy is a valuable tool that provides additional knowledge about basic metabolic processes for example in the human brain or important diagnostic information for various diseases. A topic of particular interest is the detection and quantification of neurotransmitters and their precursors such as Glutamine, Glutamate and GABA as well as antioxidants such as GSH and ascorbic acid, whose important role in the pathophysiology of psychiatric and neurological disorders is still not fully understood. Accurate quantification is essential if any conclusion should be drawn from the calculated concentrations of these metabolites. Unfortunately in a commonly used short-echo one-dimensional single voxel spectrum the independent quantification of these strongly coupled and often low concentrated resonances is hampered by the heavy overlap of the individual signals. Two-dimensional spectroscopic sequences like JRESS [1] or L-COSY [2] can be used spread out the m...

fitted Voigt line for different shim settings Figure 3: FWHMs from the spectra from all volunteer... more fitted Voigt line for different shim settings Figure 3: FWHMs from the spectra from all volunteers acquired without any shim, and with 1 and 2 order FM shim, ST shim ROI only, ST shim ROI and rect. ROLI and ST shim ROI and WH ROLI respectively Figure 1: ROI (red) and ROLIs (green) as used by the image based shim algorithm for shim calculation. The ROI is equal to the spectroscopy voxel. (a) rectangular ROLI (b) ROLI that covers the whole heart (WH ROLI)

1 Henning A., et al., Magn Reson Med, 2008;59(6):1250-8. 2. Marliani AF., et al., AJNR, 2010;31(1... more 1 Henning A., et al., Magn Reson Med, 2008;59(6):1250-8. 2. Marliani AF., et al., AJNR, 2010;31(1):180-4. 3. Dreher W, Leibfritz D. MRM, 2005;54(1):190-195. 4. de Graaf RA. In Vivo NMR Spectroscopy. Principles and Techniques. 2007 5. MacMillan EL, et al. Proc 18th ISMRM, 2010:316 6. Hock A, et al. Proc 18th ISMRM, 2010: 5042. 7. Provencher SW, et al. Magn Reson Med 1993;30:672–679. 8. S. A. Smith et. al., JMR, 1994; A 106:75–105. Figure 1. Sagittal and axial image of the cervical spinal cord. Due to the chemical shift artifact the position of the NAA (red) and mI (white) voxel is shifted. Applying IVS bands (blue) minimizes this artifact. In addition, the shim box is displayed in orange. Non-water suppressed proton MR spectroscopy allows spectral quality improvement in the human cervical spinal cord

residual (green): (a), direct acquisition; (b), with NOE and proton decoupling; (c), spectra from... more residual (green): (a), direct acquisition; (b), with NOE and proton decoupling; (c), spectra from (a) enlarged 90-120ppm; (d), spectra from (b) enlarged 90-120ppm. (a) and (b), (c) and (d) are on the same scale, respectively. The ERETIC reference appears at 115ppm, glycogen peaks are centered at 100.5ppm. Figure 1 (top) Voxel of interest in the calf muscle; (bottom) acquisition sequence combining ISIS, NOE& decoupling enhancement. ERETIC-based glycogen quantification using SNR-enhanced and localized 13C MRS

Introduction: The increase in spectral separation and signal-to-noise ratio at ultra-high field (... more Introduction: The increase in spectral separation and signal-to-noise ratio at ultra-high field (7T and higher) strength enables the detection of a large number of low concentrated or coupled spin systems aside from the big landmark peaks NAA, creatine and choline. However, their detection is complicated by relatively short T2 relaxation times at ultra-high field strength (7T). The large spectral separation and fast T2 relaxation call for a localization approach that controls the chemical shift displacement artifact while keeping echo-times short. Mlynárik et. al. [1] introduced an elegant approach combining 1D ISIS and 2D spin echo encoding for a localization scheme called SPECIAL. With this approach 3D localization can be reached by using only one slice selective refocusing pulse instead of two as in conventional localization schemes. However, besides the spatial labeling the slice selective inversion pulse used for ISIS encoding prior to the excitation also acts as a magnetizatio...

Introduction Spectral line width and signal-to-noise ratio (SNR) are crucial performance markers ... more Introduction Spectral line width and signal-to-noise ratio (SNR) are crucial performance markers for Magnetic Resonance Spectroscopic Imaging (MRSI). It is only with sufficient SNR and effective spectral resolution that reliable metabolite quantification is enabled. B0 correction at the subvoxel level [1] removes spectral shifts introduced by local static magnetic field distortions on an intermediate spectroscopic image with higher-than-nominal spatial resolution when target-driven overdiscrete reconstruction is used for the optimization of the spatial response function [2]. In this work, we show improvements in line width and SNR following this reconstruction strategy, explain the mechanisms behind the effects and demonstrate complementarity to the conventional noise reduction and resolution enhancement method of limiting the acquisition time and zero-filling the FID signal to obtain the appropriate spectral resolution. Theory and Methods Data Acquisition: A 7T MR system (Philips H...

Purpose: Magnetic resonance spectroscopy benefits greatly from increases in field strength, inclu... more Purpose: Magnetic resonance spectroscopy benefits greatly from increases in field strength, including increases in SNR and spectral separation. Thus far, the upfield part of the spectrum has been well-characterized in human brain, especially at very high field strengths [1]; however, the downfield part at 5-10ppm remains less well characterized. Some information has been published on downfield metabolites in animal brain [2, 3], and some on exchange rates and T1 values in human brain at 3T [4], but a thorough confirmation of metabolite identification and characterization is still needed, particularly in humans. This work aims to further this goal by calculating the T2 values of several peaks in the downfield spectrum in grey matter at 7 T. Materials and Methods: Spectra were acquired on a 7T Philips Achieva scanner (Philips Medical Systems, Best, The Netherlands) using a quadrature transmit/receive surface coil (Rapid Biomedical). The voxel of interest was located in the visual cortex of the brain and measured 20x40x20mm 3. Second-order local projection based B0 shimming was applied. Data was acquired using a STEAM sequence (TR/TM = 4000/24.8ms) with an eight-pulse VAPOR scheme to minimize water sidebands. A series of TEs at 13, 23, 35, 47, and 60ms was acquired in six healthy subjects (age range 21-49yrs, mean age 30yrs), with the applied RF frequency in the downfield region at 7.5ppm. For each volunteer, 256 averages per TE were collected. One data set was excluded due to excessive line broadening. The five spectra with different TEs were modeled simultaneously in FiTAID [5]. Using the average of the five remaining data sets, prior knowledge was defined with seven peaks in the 5 to 9ppm region. The NAA and α-glucose (Glc) peaks were used as binary patterns, as modeled in VESPA based on the known spin systems. Subsequently the individual data sets were fitted. Results: Fig. 1 presents the downfield region of the TE series of five volunteers summed together, with vertical lines indicating the evaluated peaks. The TE 13ms downfield spectrum, fit, and residuals from one subject are shown in Fig. 2. Average T2 results across the five subjects are given in Table 1 for the seven peaks of interest, with error given as the standard deviation across the subjects; also included are the mean and standard deviation of the Cramér-Rao bounds (CRB). Discussion: The initial fitting results indicate likely metabolite peaks at seven different locations, with fairly flat residuals (Fig. 2 shows a typical example) indicating a decent fitting result. The results for most metabolite peaks were very consistent between individuals, with low standard deviations. One exception with a very large standard deviation and CRB is the Glc T2, which proved more difficult to model because of its proximity to residual water. It should be noted that other effects such as exchange or J-coupling might affect the T2's obtained. Furthermore, for coupled peaks such as NAA, the TE decay depends critically on the TM chosen, due to zero-quantum evolution. Peak assignment, based on chemical shifts of the metabolites and also as discussed in previous work [6, 7], is somewhat difficult, particularly as the use of VAPOR for water suppression might cause peaks that exchange moderately fast to be less visible. At 3T, for example, it has been shown that exchanging peaks exist at 8.2 and 8.5ppm [4], where one would expect the rapidly exchanging amide peaks; a decreased peak intensity due to their exchange might account for the broad shoulder to the left of the NAA. Effects from macromolecules on the baseline or peaks have yet to be verified. The T2's found for the various fitted peaks are all fairly similar, but significantly shorter than those reported for upfield peaks in the brain, which are approximately 70ms or longer depending on the structure [8]. The TE dependence of the various peaks seems similar to that previously shown at 3T [9]. Although in the 3T spectra the peak at approximately 8.4ppm appears more prominent even at TE=60ms, the lower magnitude here may be due to a fielddependent T2 or more likely [4] exchange effects (longer water suppression module). Conclusions: We have recorded downfield spectra at various echo times and determined the T2 values for several peaks; while not all peaks have been identified, the quantification brings us closer to full characterization of the downfield spectrum and may aid peak assignment.

Magnetic Resonance in Medicine, 2020

This is an open access article under the terms of the Creative Commons Attribution License, which... more This is an open access article under the terms of the Creative Commons Attribution License, which permits use, distribution and reproduction in any medium, provided the original work is properly cited.

Heart, 2009

Background Patients with heart failure with preserved ejection fraction (HFpEF) have dyspnoea on ... more Background Patients with heart failure with preserved ejection fraction (HFpEF) have dyspnoea on exertion and limited exercise capacity. In this study we evaluated the role of exercise-related changes in left ventricular relaxation and of vasculoventricular coupling (VVC) as the mechanism of this limitation and whether cardiac energetic impairment may underlie these abnormalities. Methods We prospectively studied 37 patients with HFpEF. All had signs and/or symptoms of heart failure and a normal ejection fraction (>50%). Twenty age and gender-matched healthy volunteers were also studied. VVC and left ventricular filling characteristics were assessed at rest and on exercise by multiple uptake gated acquisition scan, and time to peak filling (nTTPF) was corrected for the RR interval (an indirect measure of the rate of left ventricular active relaxation) was derived. In vivo myocardial energetic state was assessed by 31P magnetic resonance spectroscopy at 3 Tesla. All subjects also ...

Introduction In vivo proton spectroscopy is a valuable tool that provides additional knowledge ab... more Introduction In vivo proton spectroscopy is a valuable tool that provides additional knowledge about basic metabolic processes for example in the human brain or important diagnostic information for various diseases. A topic of particular interest is the detection and quantification of neurotransmitters and their precursors such as Glutamine, Glutamate and GABA as well as antioxidants such as GSH and ascorbic acid, whose important role in the pathophysiology of psychiatric and neurological disorders is still not fully understood. Accurate quantification is essential if any conclusion should be drawn from the calculated concentrations of these metabolites. Unfortunately in a commonly used short-echo one-dimensional single voxel spectrum the independent quantification of these strongly coupled and often low concentrated resonances is hampered by the heavy overlap of the individual signals. Two-dimensional spectroscopic sequences like JRESS [1] or L-COSY [2] can be used spread out the m...

fitted Voigt line for different shim settings Figure 3: FWHMs from the spectra from all volunteer... more fitted Voigt line for different shim settings Figure 3: FWHMs from the spectra from all volunteers acquired without any shim, and with 1 and 2 order FM shim, ST shim ROI only, ST shim ROI and rect. ROLI and ST shim ROI and WH ROLI respectively Figure 1: ROI (red) and ROLIs (green) as used by the image based shim algorithm for shim calculation. The ROI is equal to the spectroscopy voxel. (a) rectangular ROLI (b) ROLI that covers the whole heart (WH ROLI)

1 Henning A., et al., Magn Reson Med, 2008;59(6):1250-8. 2. Marliani AF., et al., AJNR, 2010;31(1... more 1 Henning A., et al., Magn Reson Med, 2008;59(6):1250-8. 2. Marliani AF., et al., AJNR, 2010;31(1):180-4. 3. Dreher W, Leibfritz D. MRM, 2005;54(1):190-195. 4. de Graaf RA. In Vivo NMR Spectroscopy. Principles and Techniques. 2007 5. MacMillan EL, et al. Proc 18th ISMRM, 2010:316 6. Hock A, et al. Proc 18th ISMRM, 2010: 5042. 7. Provencher SW, et al. Magn Reson Med 1993;30:672–679. 8. S. A. Smith et. al., JMR, 1994; A 106:75–105. Figure 1. Sagittal and axial image of the cervical spinal cord. Due to the chemical shift artifact the position of the NAA (red) and mI (white) voxel is shifted. Applying IVS bands (blue) minimizes this artifact. In addition, the shim box is displayed in orange. Non-water suppressed proton MR spectroscopy allows spectral quality improvement in the human cervical spinal cord

residual (green): (a), direct acquisition; (b), with NOE and proton decoupling; (c), spectra from... more residual (green): (a), direct acquisition; (b), with NOE and proton decoupling; (c), spectra from (a) enlarged 90-120ppm; (d), spectra from (b) enlarged 90-120ppm. (a) and (b), (c) and (d) are on the same scale, respectively. The ERETIC reference appears at 115ppm, glycogen peaks are centered at 100.5ppm. Figure 1 (top) Voxel of interest in the calf muscle; (bottom) acquisition sequence combining ISIS, NOE& decoupling enhancement. ERETIC-based glycogen quantification using SNR-enhanced and localized 13C MRS

Introduction: The increase in spectral separation and signal-to-noise ratio at ultra-high field (... more Introduction: The increase in spectral separation and signal-to-noise ratio at ultra-high field (7T and higher) strength enables the detection of a large number of low concentrated or coupled spin systems aside from the big landmark peaks NAA, creatine and choline. However, their detection is complicated by relatively short T2 relaxation times at ultra-high field strength (7T). The large spectral separation and fast T2 relaxation call for a localization approach that controls the chemical shift displacement artifact while keeping echo-times short. Mlynárik et. al. [1] introduced an elegant approach combining 1D ISIS and 2D spin echo encoding for a localization scheme called SPECIAL. With this approach 3D localization can be reached by using only one slice selective refocusing pulse instead of two as in conventional localization schemes. However, besides the spatial labeling the slice selective inversion pulse used for ISIS encoding prior to the excitation also acts as a magnetizatio...

Introduction Spectral line width and signal-to-noise ratio (SNR) are crucial performance markers ... more Introduction Spectral line width and signal-to-noise ratio (SNR) are crucial performance markers for Magnetic Resonance Spectroscopic Imaging (MRSI). It is only with sufficient SNR and effective spectral resolution that reliable metabolite quantification is enabled. B0 correction at the subvoxel level [1] removes spectral shifts introduced by local static magnetic field distortions on an intermediate spectroscopic image with higher-than-nominal spatial resolution when target-driven overdiscrete reconstruction is used for the optimization of the spatial response function [2]. In this work, we show improvements in line width and SNR following this reconstruction strategy, explain the mechanisms behind the effects and demonstrate complementarity to the conventional noise reduction and resolution enhancement method of limiting the acquisition time and zero-filling the FID signal to obtain the appropriate spectral resolution. Theory and Methods Data Acquisition: A 7T MR system (Philips H...

Purpose: Magnetic resonance spectroscopy benefits greatly from increases in field strength, inclu... more Purpose: Magnetic resonance spectroscopy benefits greatly from increases in field strength, including increases in SNR and spectral separation. Thus far, the upfield part of the spectrum has been well-characterized in human brain, especially at very high field strengths [1]; however, the downfield part at 5-10ppm remains less well characterized. Some information has been published on downfield metabolites in animal brain [2, 3], and some on exchange rates and T1 values in human brain at 3T [4], but a thorough confirmation of metabolite identification and characterization is still needed, particularly in humans. This work aims to further this goal by calculating the T2 values of several peaks in the downfield spectrum in grey matter at 7 T. Materials and Methods: Spectra were acquired on a 7T Philips Achieva scanner (Philips Medical Systems, Best, The Netherlands) using a quadrature transmit/receive surface coil (Rapid Biomedical). The voxel of interest was located in the visual cortex of the brain and measured 20x40x20mm 3. Second-order local projection based B0 shimming was applied. Data was acquired using a STEAM sequence (TR/TM = 4000/24.8ms) with an eight-pulse VAPOR scheme to minimize water sidebands. A series of TEs at 13, 23, 35, 47, and 60ms was acquired in six healthy subjects (age range 21-49yrs, mean age 30yrs), with the applied RF frequency in the downfield region at 7.5ppm. For each volunteer, 256 averages per TE were collected. One data set was excluded due to excessive line broadening. The five spectra with different TEs were modeled simultaneously in FiTAID [5]. Using the average of the five remaining data sets, prior knowledge was defined with seven peaks in the 5 to 9ppm region. The NAA and α-glucose (Glc) peaks were used as binary patterns, as modeled in VESPA based on the known spin systems. Subsequently the individual data sets were fitted. Results: Fig. 1 presents the downfield region of the TE series of five volunteers summed together, with vertical lines indicating the evaluated peaks. The TE 13ms downfield spectrum, fit, and residuals from one subject are shown in Fig. 2. Average T2 results across the five subjects are given in Table 1 for the seven peaks of interest, with error given as the standard deviation across the subjects; also included are the mean and standard deviation of the Cramér-Rao bounds (CRB). Discussion: The initial fitting results indicate likely metabolite peaks at seven different locations, with fairly flat residuals (Fig. 2 shows a typical example) indicating a decent fitting result. The results for most metabolite peaks were very consistent between individuals, with low standard deviations. One exception with a very large standard deviation and CRB is the Glc T2, which proved more difficult to model because of its proximity to residual water. It should be noted that other effects such as exchange or J-coupling might affect the T2's obtained. Furthermore, for coupled peaks such as NAA, the TE decay depends critically on the TM chosen, due to zero-quantum evolution. Peak assignment, based on chemical shifts of the metabolites and also as discussed in previous work [6, 7], is somewhat difficult, particularly as the use of VAPOR for water suppression might cause peaks that exchange moderately fast to be less visible. At 3T, for example, it has been shown that exchanging peaks exist at 8.2 and 8.5ppm [4], where one would expect the rapidly exchanging amide peaks; a decreased peak intensity due to their exchange might account for the broad shoulder to the left of the NAA. Effects from macromolecules on the baseline or peaks have yet to be verified. The T2's found for the various fitted peaks are all fairly similar, but significantly shorter than those reported for upfield peaks in the brain, which are approximately 70ms or longer depending on the structure [8]. The TE dependence of the various peaks seems similar to that previously shown at 3T [9]. Although in the 3T spectra the peak at approximately 8.4ppm appears more prominent even at TE=60ms, the lower magnitude here may be due to a fielddependent T2 or more likely [4] exchange effects (longer water suppression module). Conclusions: We have recorded downfield spectra at various echo times and determined the T2 values for several peaks; while not all peaks have been identified, the quantification brings us closer to full characterization of the downfield spectrum and may aid peak assignment.