Navigator based respiratory gating during acquisition and preparation phases for proton liver spectroscopy at 3 T

Ectopic lipid deposition in muscle and liver, quantified by proton magnetic resonance spectroscopy

Advances in the development of noninvasive imaging techniques have spurred investigations into ectopic lipid deposition in the liver and muscle and its implications in the development of metabolic diseases such as type 2 diabetes. Computed tomography and ultrasound have been applied in the past, though magnetic resonance-based methods are currently considered the gold standard as they allow more accurate quantitative detection of ectopic lipid stores. This review focuses on methodological considerations of magnetic resonance-based methods to image hepatic and muscle fat fractions, and it emphasizes anatomical and morphological aspects and how these may influence data acquisition, analysis, and interpretation.

Fat unsaturation measures in tibial, subcutaneous and breast adipose tissue using short and long TE MRS at 3 T

Fat unsaturation and poly-unsaturation measures can be obtained in vivo with magnetic resonance spectroscopy (MRS) through the olefinic (≈5.4 ppm) and diallylic (≈2.8 ppm) resonances, respectively. Long echo time (TE) MRS sequences have been previously optimized for olefinic/methylene (≈1.3 ppm) or olefinic/methyl (≈0.9 ppm) measures. The objectives of this work, using a Point RESolved Spectroscopy (PRESS) sequence, are to: 1) Investigate olefinic, methyl and methylene resonance decay in subcutaneous, tibial, and breast adipose tissue to determine if a direct comparison of unsaturation measures can be made without correction for T₂ losses. 2) Assess intra-individual fat unsaturation and poly-unsaturation measures in the three adipose tissues. 3) Estimate correction factors for olefinic to methylene ratios to compensate for J-coupling and T₂ relaxation losses that take place when increasing PRESS TE from 40 ms to 200 ms (previously optimized long-TE). 4) Investigate the utility of an inversion recovery for resolving the olefinic resonance from water in adipose tissue. PRESS spectra were acquired from the three adipose regions (breast in female only) in healthy volunteers at 3 T. It was found that olefinic and methyl signal decays faster in breast tissue compared to in tibial bone marrow. Poly-unsaturation measures (diallylic/methylene) differ for tibial bone marrow compared to subcutaneous and breast adipose tissue, with average values of 1.7 ± 0.4, 2.2 ± 0.4, and 2.3 ± 0.8%, respectively. PRESS (TE = 40 ms) with an inversion recovery resolves the olefinic and water resonances in breast tissue with a signal to noise ratio approximately six times greater than that using PRESS with a TE of 200 ms. Stimulated Echo Acquisition Mode (STEAM) with a TE of 20 ms (mixing time of 20 ms) was also combined with IR to resolve the olefinic resonance from that of water is spinal bone marrow.

B0 shimming for in vivo magnetic resonance spectroscopy: Experts' consensus recommendations

Magnetic resonance spectroscopy (MRS) and spectroscopic imaging (MRSI) allow the chemical analysis of physiological processes in vivo and provide powerful tools in the life sciences and for clinical diagnostics. Excellent homogeneity of the static B₀ magnetic field over the object of interest is essential for achieving high-quality spectral results and quantitative metabolic measurements. The experimental minimization of B₀ variation is performed in a process called B₀ shimming. In this article, we summarize the concepts of B₀ field shimming using spherical harmonic shimming techniques, specific strategies for B₀ homogenization and crucial factors to consider for implementation and use in both brain and body. In addition, experts' recommendations are provided for minimum requirements for B₀ shim hardware and evaluation criteria for the primary outcome of adequate B₀ shimming for MRS and MRSI, such as the water spectroscopic linewidth.

Levels of choline-containing compounds in normal liver and liver metastases of colorectal cancer as recorded by 1 H MRS

PURPOSE

A relatively high signal for choline-containing compounds (total choline, tCho) is commonly found in ¹ H MR spectra of malignant tumors, but it is unclear if this also occurs in tumors in the liver. We evaluated the potential of the tCho signal in single voxel ¹ H MR spectra of the human liver to assess metastases of colorectal cancers.

EXPERIMENT

MR spectra of an 8 cm³ PRESS-localized voxel were obtained at 3 T from the livers of 12 healthy volunteers and from metastatic lesions in 20 patients in two different sessions. To correct for motion artifacts, sequentially recorded spectra were individually phased and frequency aligned before averaging. Spectra were analyzed using LCModel and tissue levels estimated by water referencing. Repeatability was assessed with Bland-Altman analyses. To estimate tumor necrosis, diffusion-weighted imaging of the liver was performed. High resolution magic angle spinning (HRMAS) spectra of tumor and normal liver samples were obtained at 11.7 T.

RESULTS

With increasing tumor volumes, tCho levels decreased, indicating a partial volume effect. Mean tCho content in tumors larger than the PRESS voxel (>8 cm³ ) was significantly lower (p < 0.01) than for normal liver: 1.6 (range 0.0-3.4) versus 6.9 (range 4.9-11.1) mmol/kg wet weight, while it was comparable for tumors smaller than 8 cm³ : 7.0 (range 3.8-9.3) mmol/kg. The higher 90th percentile apparent diffusion coefficient value in the larger lesions indicates more necrosis. Measurement repeatability was average in normal livers and poor in tumors. HRMAS did not show substantial differences in choline-containing compounds between normal liver and metastasis.

CONCLUSION

An increased tCho content was not observed in ¹ H MR spectra of liver metastasis of colorectal cancer, compared with normal liver. This may be due to the background of a high tCho signal in spectra of normal liver or to an intrinsic lower tCho content in these tumors, but is most likely the result of necrosis in metastatic tumor tissue.

Non-Water-Suppressed 1H MR Spectroscopy with Orientational Prior Knowledge Shows Potential for Separating Intra- and Extramyocellular Lipid Signals in Human Myocardium

Conditions such as type II diabetes are linked with elevated lipid levels in the heart, and significantly increased risk of heart failure; however, metabolic processes underlying the development of cardiac disease in type II diabetes are not fully understood. Here we present a non-invasive method for in vivo investigation of cardiac lipid metabolism: namely, IVS-McPRESS. This technique uses metabolite-cycled, non-water suppressed 1H cardiac magnetic resonance spectroscopy with prospective and retrospective motion correction. High-quality IVS-McPRESS data acquired from healthy volunteers allowed us to investigate the frequency shift of extramyocellular lipid signals, which depends on the myocardial fibre orientation. Assuming consistent voxel positioning relative to myofibres, the myofibre angle with the magnetic field was derived from the voxel orientation. For separation and individual analysis of intra- and extramyocellular lipid signals, the angle myocardial fibres in the spectroscopy voxel take with the magnetic field should be within ±24.5°. Metabolite and lipid concentrations were analysed with respect to BMI. Significant correlations between BMI and unsaturated fatty acids in intramyocellular lipids, and methylene groups in extramyocellular lipids were found. The proposed IVS-McPRESS technique enables non-invasive investigation of cardiac lipid metabolism and may thus be a useful tool to study healthy and pathological conditions.

Fully navigated 3 T proton magnetic resonance spectroscopy of liver metastases with inner-volume saturation

PURPOSE

To demonstrate that fully navigated magnetic resonance spectroscopy (MRS) with inner-volume saturation (IVS) at 3 T results in high-quality spectra that permit evaluating metabolic changes in hepatic metastases without the need for patient compliance.

METHODS

Nine patients with untreated, biopsy-proven large hepatic metastases (minimum diameter of 3 cm) were included. In each patient, localized proton MRS was performed in the metastatic lesion and in uninvolved liver parenchyma. To improve quality and consistency of proton MRS, navigator gating was thereby performed not only during acquisition of the spectroscopic data but also during localization imaging and throughout the preparation phases. IVS was utilized to reduce chemical shift displacement between different metabolites and to diminish flow artifacts. Metabolite quantities were normalized relative to the unsuppressed water peak and choline-containing compounds (CCC) to lipid ratios were determined. Wilcoxon signed-rank tests were used to assess differences in the amounts of lipids and CCC as well as the CCC-to-lipid ratios between liver metastases and normal-appearing liver parenchyma.

RESULTS

Fully navigated point-resolved spectroscopy with IVS resulted in high-quality spectra in all patients. Navigator gating during localization imaging and spectroscopic acquisition thereby ensured a precise localization of the spectroscopic voxel. Decreased quantities of lipid and CCC were observed in metastatic tissue compared with uninvolved liver parenchyma. However, the latter trend fell short of statistical significance. Moreover, elevated levels of the CCC-to-lipid ratios were detected in metastatic tissue relative to normal-appearing liver parenchyma.

CONCLUSIONS

The present study demonstrates that fully navigated MRS of the liver with IVS at 3 T allows for a precise localization of the spectroscopic voxel and results in high-quality spectra that permit evaluating liver metabolism without the need for patient compliance.

Multiple breath-hold proton spectroscopy of human liver at 3T: Relaxation times and concentrations of glycogen, choline, and lipids

PURPOSE

To evaluate the feasibility of an expiration multiple breath-hold ¹ H-MRS technique to measure glycogen (Glycg), choline-containing compounds (CCC), and lipid relaxation times T₁ , T₂ , and their concentrations in normal human liver.

MATERIALS AND METHODS

Thirty healthy volunteers were recruited. Experiments were performed at 3T. Multiple expiration breath-hold single-voxel point-resolved spectroscopy (PRESS) technique was used for localization. Water-suppressed spectra were used for the estimation of Glycg, CCC, lipid methylene (CH₂ )_n relaxation times and concentrations. Residual water lines were removed by the Hankel Lanczos singular value decomposition filter. After phase correction and frequency alignment, spectra were averaged and processed by LCModel. Summed signals of Glycg resonances H2H4', H3, and H5 between 3.6 and 4 ppm were used to estimate their apparent relaxation times and concentration. Glycg, CCC, and lipid content were estimated from relaxation corrected spectral intensity ratios to unsuppressed water line.

RESULTS

Relaxation times were measured for liver Glycg (T₁ , 892 ± 126 msec; T₂ , 13 ± 4 msec), CCC (T₁ , 842 ± 75 msec; T₂ , 50 ± 5 msec), lipid (CH₂ )_n (T₁ , 402 ± 19 msec; T₂ , 52 ± 3 msec), and water (T₁ , 990 ± 89 msec; T₂ , 30 ± 2 msec). Mean CCC and lipid concentrations of healthy liver were 7.8 ± 1.3 mM and 15.8 ± 23.6 mM, respectively. Glycg content was found lower in the morning (48 ± 21 mM) compared to the afternoon (145 ± 50 mM).

CONCLUSION

Multiple breath-hold ¹ H-MRS together with dedicated postprocessing is a feasible technique for the quantification of liver Glycg, CCC, and lipid relaxation times and concentrations.

LEVEL OF EVIDENCE

1 Technical Efficacy: Stage 2 J. Magn. Reson. Imaging 2018;47:410-417.