Outer Volume Suppression Research Articles

In vivo metabolite profile of adult zebrafish brain obtained by high‐resolution localized magnetic resonance spectroscopy

To optimize high-resolution MR spectroscopy (MRS) for obtaining neurochemical composition of adult zebrafish brain in vivo. A flow-through setup for supporting MRS of living zebrafish has been designed. In vivo MR microscopy (MRM) images were obtained using a rapid acquisition with relaxation enhancement (RARE) sequence to select a volume of interest. In vivo MR spectra from zebrafish brain were obtained using an optimized point-resolved spectroscopy (PRESS) sequence preceded by a variable pulse power and optimized relaxation delays (VAPOR) sequence for global water suppression interleaved with outer volume suppression (OVS). In vitro MR spectra in the brain extract were obtained by using correlated spectroscopy (COSY) sequences. Optimized high-resolution localized MRS at 9.4T in conjunction with a strong gradient system, efficient shimming, and the water suppression scheme resulted in a reasonable separation of resonances from various metabolites in vivo from a voxel as small as 3.3 microL placed in the zebrafish brain. In addition, more than 14 metabolites were identified in adult zebrafish brain extracts. We have successfully optimized a high-resolution localized in vivo MRS technique to get access to the zebrafish brain, and obtained for the first time the neurochemical composition of the zebrafish brain.

Diffusion‐weighted imaging of the entire spinal cord

In spite of their diagnostic potential, the poor quality of available diffusion-weighted spinal cord images often restricts clinical application to cervical regions, and improved spatial resolution is highly desirable. To address these needs, a novel technique based on the combination of two recently presented reduced field-of-view approaches is proposed, enabling high-resolution acquisition over the entire spinal cord. Field-of-view reduction is achieved by the application of non-coplanar excitation and refocusing pulses combined with outer volume suppression for removal of unwanted transition zones. The non-coplanar excitation is performed such that a gap-less volume is acquired in a dedicated interleaved slice order within two repetition times. The resulting inner volume selectivity was evaluated in vitro. In vivo diffusion tensor imaging data on the cervical, thoracic and lumbar spinal cord were acquired in transverse orientation in each of four healthy subjects. An in-plane resolution of 0.7 x 0.7 mm(2) was achieved without notable aliasing, motion or susceptibility artifacts. The measured mean +/- SD fractional anisotropy was 0.69 +/- 0.11 in the thoracic spinal cord and 0.75 +/- 0.07 and 0.63 +/- 0.08 in cervical and lumbar white matter, respectively.

Quantitative magnetic resonance spectroscopy in the entire human cervical spinal cord and beyond at 3T

Quantitative magnetic resonance spectroscopy (MRS) amends differential diagnostics of neurological pathology. However, due to technical challenges, it has rarely been applied to the spinal cord and has mainly been restricted to the very upper part of the cervical spine. In this work, an improved acquisition protocol is proposed that takes technical problems as strong magnetic field inhomogeneities, pulsatile flow of the cerebrospinal fluid (CSF), and small voxel size into account. For that purpose, inner-volume saturated point-resolved spectroscopy sequence (PRESS) localization, ECG triggering, and localized higher-order shimming and F0 determination, based on high-resolution cardiac-triggered static magnetic field B0 mapping, are combined. For inner-volume saturation a highly selective T1- and B1-insensitive outer-volume suppression (OVS) sequence based on broadband RF pulses with polynomial-phase response (PPR) is used. Validation is performed in healthy volunteers and patients with multiple sclerosis and intramedullary tumors. The applicability of spinal cord MRS is extended to the entire cervical spine. Spectral quality and its consistency are improved. In addition, high quality MRS patient data from a lesion that occluded the spinal canal in the thoracic spinal cord could be acquired. A quantitative analysis of patient spectra and spectra from healthy volunteers at different positions along the spinal cord underlines the diagnostic value of spinal cord MRS.

SELOVS: Brain MRSI localization based on highly selective T1‐ and B1‐insensitive outer‐volume suppression at 3T

In vivo, high-field MR spectroscopic imaging (MRSI) profits from signal-to-noise ratio (SNR) gain and increased spectral resolution. However, bandwidth limitations of slice-selective excitation and refocusing pulses lead to strong chemical-shift displacement at high field strength when using conventional MRSI localization based on PRESS. Consequential metabolic information, particularly of border regions such as cortical brain tissue, is distorted. In addition, lipid contamination remains a major confound. To address these problems it is proposed to abandon PRESS selection and rely on a novel scheme of highly selective T(1)- and B(1)-insensitive outer-volume suppression in combination with slice-selective spin-echo acquisition for brain MRSI. Multiple cycles of overlapping suppression slabs are applied with flip angles optimized to account for tissue-dependent T(1) relaxation times and band crossings. Broadband frequency modulated saturation pulses with polynomial phase-response are utilized in order to minimize chemical-shift displacement. Efficacy of the outer-volume suppression sequence was simulated and evaluated in vitro and in vivo. Brain MRSI localization at 3T was significantly improved and reliable suppression of short-range lipid contamination enabled, leading to substantial enhancement of spectral quality, particularly in cortical tissue. Hence, the new method holds potential to expand the applicability of high-field MRSI to the entire brain.

Proton MR spectroscopy of the prostate

Purpose To summarize current technical and biochemical aspects and clinical applications of proton magnetic resonance spectroscopy (MRS) of the human prostate in vivo. Material and methods Pertinent radiological and biochemical literature was searched and retrieved via electronic media (medline ®, pubmed ®). Basic concepts of MRS of the prostate and its clinical applications were extracted. Results Clinical MRS is usually based on point resolved spectroscopy (PRESS) or spin echo (SE) sequences, along with outer volume suppression of signals from outside of the prostate. MRS of the prostate detects indicator lines of citrate, choline, and creatine. While healthy prostate tissue demonstrates high levels of citrate and low levels of choline that marks cell wall turnover, prostate cancer utilizes citrate for energy metabolism and shows high levels of choline. The ratio of (choline + creatine)/citrate distinguishes between healthy tissue and prostate cancer. Particularly when combined with magnetic resonance (MR) imaging, three-dimensional MRS imaging (3D-CSI, or 3D-MRSI) detects and localizes prostate cancer in the entire prostate with high sensitivity and specificity. Combined MR imaging and 3D-MRSI exceed the sensitivity and specificity of sextant biopsy of the prostate. When MRS and MR imaging agree on prostate cancer presence, the positive predictive value is about 80–90%. Distinction between healthy tissue and prostate cancer principally is maintained after various therapeutic treatments, including hormone ablation therapy, radiation therapy, and cryotherapy of the prostate. Conclusions Since it is non-invasive, reliable, radiation-free, and essentially repeatable, combined MR imaging and 3D-MRSI of the prostate lends itself to the planning of biopsy and therapy, and to post-therapeutic follow-up. For broad clinical acceptance, it will be necessary to facilitate MRS examinations and their evaluation and make MRS available to a wider range of institutions.

Proton echo‐planar spectroscopic imaging of J‐coupled resonances in human brain at 3 and 4 Tesla

In this multicenter study, 2D spatial mapping of J-coupled resonances at 3T and 4T was performed using short-TE (15 ms) proton echo-planar spectroscopic imaging (PEPSI). Water-suppressed (WS) data were acquired in 8.5 min with 1-cm(3) spatial resolution from a supraventricular axial slice. Optimized outer volume suppression (OVS) enabled mapping in close proximity to peripheral scalp regions. Constrained spectral fitting in reference to a non-WS (NWS) scan was performed with LCModel using correction for relaxation attenuation and partial-volume effects. The concentrations of total choline (tCho), creatine + phosphocreatine (Cr+PCr), glutamate (Glu), glutamate + glutamine (Glu+Gln), myo-inositol (Ins), NAA, NAA+NAAG, and two macromolecular resonances at 0.9 and 2.0 ppm were mapped with mean Cramer-Rao lower bounds (CRLBs) between 6% and 18% and approximately 150-cm(3) sensitive volumes. Aspartate, GABA, glutamine (Gln), glutathione (GSH), phosphoethanolamine (PE), and macromolecules (MMs) at 1.2 ppm were also mapped, although with larger mean CRLBs between 30% and 44%. The CRLBs at 4T were 19% lower on average as compared to 3T, consistent with a higher signal-to-noise ratio (SNR) and increased spectral resolution. Metabolite concentrations were in the ranges reported in previous studies. Glu concentration was significantly higher in gray matter (GM) compared to white matter (WM), as anticipated. The short acquisition time makes this methodology suitable for clinical studies.

TH-C-L100J-05: In Vivo Prostate MRSI Using An Improved Outer Volume Suppression Technique

Purpose: To implement, in vivo, an optimized prostate 3D MR SpectroscopyImaging(MRSI) technique which improves outer volume suppression of peripheral lipid and reduces lipid contamination of prostate spectroscopicimaging.Method and Materials: All scans were performed on a General Electric 1.5T Signa MR scanner equipped with Echospeed gradients. A standard quadrature head coil was used for all phantom experiments and an endorectal coil (Medrad Inc.) in combination with a torso phased arraycoil was used for all human experiments. A water/oil phantom was designed to simulate the prostate and surrounding lipid signal. The product spectroscopy pulse sequence was modified to include twenty or more optimized spatial saturation pulses. 3D MRSI spectra, employing the optimized prostate spectroscopy technique, was collected from both phantom and in vivo prostate. Results: Phantom results obtained using a standard 3D MRSI technique, which uses 4 manually placed spatial saturation pulses, showed significant lipid contamination well within the object where lipid is not present. Using the optimized spectroscopic technique on the prostate phantom we observed a 80% reduction in peripheral lipid. The 3D MRSI spectra revealed that significant reduction within all parts of the phantom were observed. Consistent with our phantom experiments, initial in vivo 3D MRSI of the prostate demonstrates, in some voxels, up to 100% reduction of lipid contamination due to peripheral contamination. Conclusion: Phantom results indicate that significant lipid contamination occurs when using manually placed spatial saturation pulses. Using this novel, optimized MRSI technique has significantly reduced the problem related to lipid contamination. In vivo, implementation of the optimized MRSI technique has confirmed the decrease in peripheral lipid contamination. Thus a technique has been established which reduces the negative effect of perisprostatic lipid in prostate spectroscopy.

Refinement of cardiac NMR imaging in awake hamsters: proof of feasibility and characterization of cardiomyopathy

The goal of this study was to demonstrate the feasibility of cardiac NMR imaging in conscious hamsters and its usefulness in evaluating cardiac abnormalities in a small-animal model of cardiomyopathy. Awake hamsters, controls and cardiomyopathic ones (CHF 147), were immobilized in a dedicated holder. Half-Fourier single-shot FSE imaging, with outer-volume suppression and 'black-blood' contrast provided images free from motion artifact with good visualization of cardiac anatomy at any point in the cardiac cycle. Series of double-oblique views were acquired with or without electrocardiograph gating. Image acquisition time was 55 ms, with an in-plane resolution of 470 x 625 microm2. Left ventricular volumes, ejection fraction, and myocardium NMR signal heterogeneity were compared in CHF 147 and control hearts. Left ventricles of CHF 147 hamsters were dilated, as indicated by the increase in end-diastolic cavity volume (299 +/- 79 mm3 compared with the controls (141 +/- 39 mm3; P = 0.0002). Left ventricular ejection fraction was largely reduced (45 +/- 9% vs 86 +/- 4%; P < 0.0001). The NMR signal distribution at an effective echo time of 41 ms was more heterogeneous in the myocardial wall of CHF 147 hamsters than in controls (1.87 +/- 0.37 a.u. vs 0.98 +/- 0.12 a.u., respectively; P = 0.0002). This study is a refinement of animal experimentation, as it demonstrates for the first time that characteristic features of cardiac pathology can be evaluated with ultra-fast NMR imaging in conscious small rodents.

Reduced field‐of‐view MRI using outer volume suppression for spinal cord diffusion imaging

A spin-echo single-shot echo-planar imaging (SS-EPI) technique with a reduced field of view (FOV) in the phase-encoding direction is presented that simultaneously reduces susceptibility effects and motion artifacts in diffusion-weighted (DW) imaging (DWI) of the spinal cord at a high field strength (3T). To minimize aliasing, an outer volume suppression (OVS) sequence was implemented. Effective fat suppression was achieved with the use of a slice-selection gradient-reversal technique. The OVS was optimized by numerical simulations with respect to T(1) relaxation times and B(1) variations. The optimized sequence was evaluated in vitro and in vivo. In simulations the optimized OVS showed suppression to <0.25% and approximately 3% in an optimal and worst-case scenario, respectively. In vitro measurements showed a mean residual signal of <0.95% +/- 0.42 for all suppressed areas. In vivo acquisition with 0.9 x 1.05 mm(2) in-plane resolution resulted in artifact-free images. The short imaging time of this technique makes it promising for clinical studies.

Restricted field of view magnetic resonance imaging of a dynamic time series

A restricted field of view (rFOV) approach for imaging a dynamic time series of volumes of limited spatial extent within a larger subject is described. The shorter readout with rFOV-MRI can be exploited to either limit image artifacts or increase spatial resolution. To accomplish rFOV imaging of a multislice volume for a dynamic series, an outer volume suppression (OVS) preparation that saturates signal external to a cylinder through the subject is followed by slice-selective excitation and a spiral readout. The pass- and stopband efficiencies of the OVS in an agar gel phantom were 97% (+/-1.5%) and 3% (+/-1%), respectively. Profiles of the temporal signal-to-noise ratio (SNR) were measured in a phantom and an adult brain. The rFOV sequence reduced distortions from off-resonance signal and T2*-induced blurring compared to a conventional sequence. Sequence utility is demonstrated for high-resolution rFOV functional MRI (fMRI) in the visual cortex. The rFOV sequence may prove to be useful for other multislice dynamic and high-resolution imaging applications.

Improved echo volumar imaging (EVI) for functional MRI

Echo volumar imaging (EVI) is a 3D modification of echo-planar imaging (EPI) that allows data from an entire volume to be acquired following a single RF excitation. EVI provides a high rate of volumar data acquisition, which is advantageous for functional MRI (fMRI). However, few studies to date have applied EVI to fMRI, since because of gradient hardware limitations EVI generally has to be used with long sampling times, resulting in high sensitivity to susceptibility-induced distortions. In this study we modified the EVI sequence to improve its suitability for fMRI. The sampling time is reduced by the use of a high gradient-switching frequency, a small number of echoes, and outer volume suppression (OVS); rewind gradients ameliorate Nyquist ghosting; and phase correction via a calibration scan reduces ghosting and distortion. It is shown that the modified EVI sequence allows fMRI data to be acquired with a temporal resolution of 167 ms.

In vivo dynamic turnover of cerebral 13C isotopomers from [U– 13C]glucose

An INEPT-based 13C MRS method and a cost-effective and widely available 11.7 Tesla 89-mm bore vertical magnet were used to detect dynamic 13C isotopomer turnover from intravenously infused [U– 13C]glucose in a 211 μL voxel located in the adult rat brain. The INEPT-based 1H → 13C polarization transfer method is mostly adiabatic and therefore minimizes signal loss due to B 1 inhomogeneity of the surface coils used. High quality and reproducible data were acquired as a result of combined use of outer volume suppression, ISIS, and the single-shot three-dimensional localization scheme built in the INEPT pulse sequence. Isotopomer patterns of both glutamate C4 at 34.00 ppm and glutamine C4 at 31.38 ppm are dominated first by a doublet originated from labeling at C4 and C5 but not at C3 (with 1 J C4C5 = 51 Hz) and then by a quartet originated from labeling at C3, C4, and C5 (with 1 J C3C4 = 35 Hz). A lag in the transition of glutamine C4 pattern from doublet-dominance to quartet dominance as compared to glutamate C4 was observed, which provides an independent verification of the precursor–product relationship between neuronal glutamate and glial glutamine and a significant intercompartmental cerebral glutamate–glutamine cycle between neurons and glial cells.

TU‐C‐330A‐05: Optimization of Outer Volume Suppression for Improved Prostate MR Spectroscopic Imaging

Purpose: To adapt a new MR Spectroscopy (MRS) technique employing non‐cuboidal voxels, called conformal voxel MRS (CV‐MRS), for use in prostate spectroscopic imaging in order to reduce contamination of spectra by lipid signal surrounding the prostate.Method and Materials: CV‐MRS uses twenty or more spatial saturation (SS) pulses, placed around the prostate, to reduce the lipid signal affecting the spectra within the prostate. A water/oil phantom was designed to simulate the prostate and surrounding lipid signal. Use of the new CV‐MRS technique reduced the lipid signal contamination by 84% as compared to standard cuboidal voxel MRS. To further reduce the lipid contamination, the routinely used 90 degree flip angle used for each SS pulse was modified to take into account the regrowth of lipid signal with its short T1 relaxation time. Results: Contrary to our expectations, resulting spectra from the optimized approach actually showed an increase in lipid contamination by 10%. We tracked the problem down to overlapping SS pulses. Using a simulated 3D model, we found that 68% of the volume we were trying to saturate experienced multiple overlapping SS pulses, with some regions being saturated 7 or more times. Regions of the volume experiencing an even number of SS pulses were found to increase the lipid contamination signal by 88% to 200%. Conversely, regions experiencing an odd number of SS pulses had a reduction in lipid contamination of 55%. Conclusion: Changing the ordering of the SS pulses, such that the overlapping pulses occur later in the train of 20 SS pulses reduced the problem of lipid signal from those overlapping volumes. In summary, we have developed an improved outer volume saturation technique which reduces lipid contamination problems in prostate MR spectroscopic imaging.

Simultaneous outer volume and blood suppression by quadruple inversion‐recovery

A new method has been developed for reduced field-of-view (FOV) imaging with simultaneous blood suppression. This method combines suppression of signals from the outer volume and inflowing blood by using a small-FOV quadruple inversion-recovery (SFQIR) preparative pulse sequence consisting of two double-inversion pulse pairs separated by appropriate delays. Within each pair, inversion pulses are successively applied to the imaged slice and the slab orthogonal to the imaging plane with the thickness equal to the FOV size in the phase-encoding direction. Each double inversion results in the reinversion of the magnetization in the central part of the FOV, while the outer areas of the FOV and inflowing blood remain inverted. The SFQIR module was implemented for single- and multislice acquisition with a fast spin-echo readout sequence. Based on a theoretical model of the signal, the timing parameters of the sequence corresponding to the maximal suppression efficiency can be found by minimizing the variation of the normalized signal over the entire range of T1's that occur in tissues. The method was tested for black-blood imaging of the aorta and carotid arteries, and the results demonstrated its ability to eliminate motion and flow artifacts, reduce scan time, and improve spatial resolution.

Ultra-short echo time spectroscopic imaging in rats: implications for monitoring lipids in glioma gene therapy

We demonstrate the feasibility of using ultra-short echo time (TE = 2 ms) magnetic resonance spectroscopic imaging (MRSI) to detect intracranial mobile lipids in the rat brain. High-performance outer volume suppression and pre-localization were demonstrated in phantoms and by the total absence of signals arising from extra-cranial lipids in MRSI spectra from control rats. The sequence performance was tested on glioma-bearing BDIX rats. Fast-relaxing lipid signals were spatially varied within a glioma during herpes simplex virus thymidine kinase-mediated gene therapy, demonstrating the potential application of this method.

SU‐FF‐I‐31: An Improved Method for Susceptibility Correction of MR Spectroscopic Images

Purpose: Multivoxel spectroscopic imaging usually requires a correction for magnetic field differences between voxels. The correction enables a narrower frequency range for integration to produce the metabolic image, increasing the signal to noise ratio of the image. One method employs a single reference peak common to spectra in all voxels, such as n‐acetylaspartate, NAA. This works well in all but tumor cases where NAA is not present. A manual search must be done to find and correct the spectra in those voxels by referring to the creatine or choline peaks. This is rather tedious so a method of using multiple reference peaks was developed. If the pattern of the reference peaks is recognized, it can be used to correct for the magnetic field, automatically. Method and Materials: Data was acquired with the sequence from the NIH on 1.5T and 3T magnets from GE Healthcare. The sequence includes water saturation and outer volume suppression and allows for 4 slices to be acquired with a 32×32 matrix yielding a voxel size of 0.9 cc. The analysis was performed on Eigentool software with spectroscopy analysis code CSX. The reference peaks are defined by with a window and a maximum. The pattern of space between the reference peaks is compared to that of peaks larger than a second threshold. The pattern is allowed to shift up to one quarter of the span between the reference peaks. Results: The technique increases the speed of the analysis by decreasing the number of voxels that need to be checked manually from 100 or more to just the perimeter, perhaps 25, where the outer volume suppression may have missed some fat. Conclusion: Improving the analysis of MRSI can improve patient care by speeding up the time to report the results to the referring physician.

Methodology of1H NMR spectroscopy of the human brain at very high magnetic fields

An ultrashort-echo-time stimulated echo-acquisition mode (STEAM) pulse sequence with interleaved outer volume suppression and VAPOR (variable power and optimized relaxation delays) water suppression was redesigned and optimized for human applications at 4 and 7 T, taking into account the specific requirements for spectroscopy at high magnetic fields and limitations of currently available hardware. In combination with automatic shimming, automated parameter adjustments and data processing, this method provided a user-friendly tool for routine (1)H nuclear magnetic resonance (NMR) spectroscopy of the human brain at very high magnetic fields. Effects of first- and second-order shimming, single-scan averaging, frequency and phase corrections, and eddy currents were described. LCModel analysis of an in vivo (1)H NMR spectrum measured from the human brain at 7 T allowed reliable quantification of more than fifteen metabolites noninvasively, illustrating the potential of high-field NMR spectroscopy. Examples of spectroscopic studies performed at 4 and 7 T demonstrated the high reproducibility of acquired spectra quality.

A localization method for the measurement of fast relaxing13C NMR signals in humans at high magnetic fields

Nuclear magnetic resonance (NMR) signals with shortT 1 andT 2, such as the13C signal of glycogen, are difficult to localize in three dimensions without major signal loss. A pulse sequence that accomplishes the spatial localization of1H-decoupled13C NMR signals on a whole-body scanner within the Food and Drug Administration guidelines for specific absorption rates was designed. The method uses an optimized three-dimensional outer volume suppression scheme combined with one-dimensional image-selected in vivo spectroscopy and surface coil detection. The localization performance of the sequence was validated at 4 T with double chambered phantoms and13C magnetic resonance imaging. Localized13C spectra were acquired from human brain and muscle.

Bandwidth‐modulated adiabatic RF pulses for uniform selective saturation and inversion

Radiofrequency (RF) inversion and saturation pulses with extremely high spatial selectivity and uniform profiles are a requirement for numerous MR techniques, such as pulsed arterial spin labeling and outer volume suppression. Adiabatic pulses used for inversion of longitudinal magnetization are ubiquitous, but the superior selectivity of adiabatic full passages has not been widely exploited for saturation because a simple way of calibrating the amplitude of these subadiabatic pulses is lacking. An analytically derived calibration equation is presented, applicable to a large class of pulses including the hyperbolic secant (HS) pulse and allowing the determination of the precise amplitude required to achieve any effective flip angle. The properties of this calibration are examined, and a highly selective and homogeneous HS saturation pulse is demonstrated. Based on this calibration a new class of RF pulses is developed. These bandwidth-modulated adiabatic selective saturation and inversion (BASSI) RF pulses afford optimal amplitude modulation, achieving uniform profiles at any effective flip angle. BASSI pulses are compared to existing gradient modulated adiabatic pulses in simulations and phantom experiments and shown to be superior in terms of selectivity and homogeneity, while requiring less RF energy. An application of BASSI pulses to pulsed arterial spin labeling is shown.

LIVER GLYCOGEN AND LIVER VOLUME CHANGES WITH EXERCISE

The contribution of the liver to glucose provision during exercise is less well studied than that of muscle. Few studies have measured liver glycogen and volume changes as a result of exercise in trained subjects. PURPOSE To measure liver glycogen and liver volume changes as a result of long-lasting exercise. METHODS Six trained males (mean ± SD age 35.8 ± 7.3 yrs, mass 73.5 ± 5.1 kg, VO2max 60.8 ± 6.4 ml·min-1.kg-1) cycled 120 min (one subject 107 min) at 65% VO2max. Standardised evening meal, breakfast and snack (2 hr before testing) were eaten, and no exercise was done 24 h, before testing. Immediately before and after exercise liver volume and glycogen were measured. Glycogen was measured with MRS (Siemens 1.5T). To distinguish liver from muscle, outer volume suppression was used and chemical shift imaging was used to select a voxel of liver tissue of 125 ml. The coil configuration consisted of a 1H coil (15 cm diameter) for decoupling and a 13C surface coil (13 cm diameter) for acquisition. Volume was measured with multi-slice MRI with the 1H coil. Blood samples were taken for glucose and insulin concentration at rest and every 30 min during, and upon completion of, exercise. RESULTS Liver glycogen decreased by 49 ± 30%. Two subjects had extremely low glycogen use (14, 17%). One of these two subjects had higher resting insulin concentration (17 versus range others 6–9 mlU.L-1) and both had higher exercise insulin concentration (mean of 4 samples during exercise: 7 and 8 versus range others 1.8 – 3.3 mIU.L-1). Liver volume (n=5) deceased from 1.74 ± 0.1 to 1.52 ± 0.05 L, i.e. by 12.6 ± 2.4% CONCLUSION Decreases in liver glycogen concentration with 65% VO2max cycling are in line with the limited data published to date with submaximal exercise to exhaustion. An extremely large variation in liver glycogen use was observed. It is probable that non-compliance to the prescribed eating, as indicated by high insulin values and possibly, in one subject, altered glucose regulation and insulin production account for the extreme variability.