| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
|
|
|
|||||||
|
|||||||||
LETTER TO THE EDITOR
The following is the abstract of the article discussed in the subsequent letter:
Although skeletal muscle perfusion is fundamental to proper muscle function, in vivo measurements are typically limited to those of limb or arterial blood flow, rather than flow within the muscle bed itself. We present a noninvasive functional MRI (fMRI) technique for measuring perfusion-related signal intensity (SI) changes in human skeletal muscle during and after contractions and demonstrate its application to the question of occlusion during a range of contraction intensities. Eight healthy men (aged 20–31 yr) performed a series of isometric ankle dorsiflexor contractions from 10 to 100% maximal voluntary contraction. Axial gradient-echo echo-planar images (repetition time = 500 ms, echo time = 18.6 ms) were acquired continuously before, during, and following each 10-s contraction, with 4.5-min rest between contractions. Average SI in the dorsiflexor muscles was calculated for all 240 images in each contraction series. Postcontraction hyperemia for each force level was determined as peak change in SI after contraction, which was then scaled to that obtained following a 5-min cuff occlusion of the thigh (i.e., maximal hyperemia). A subset of subjects (n = 4) performed parallel studies using venous occlusion plethysmography to measure limb blood flow. Hyperemia measured by fMRI and plethysmography demonstrated good agreement. Postcontraction hyperemia measured by fMRI scaled with contraction intensity up to 60% maximal voluntary contraction. fMRI provides a noninvasive means of quantifying perfusion-related changes during and following skeletal muscle contractions in humans. Temporal changes in perfusion can be observed, as can the heterogeneity of perfusion across the muscle bed.
We commend D. M. Wigmore et al. (7) for having studied in vivo the influence of isometric muscle contractions upon muscle water signal. They used a T1-weighted sequence, sensitive to perfusion-induced T1 variations. This gradient-echo sequence at echo time (TE) = 18.6 ms is also sensitive to T2* variations and then reflects variations of blood oxygenation known as blood oxygen-dependent (BOLD) contrast (4). Because the variations of blood oxygenation reveal a modified balance between oxygen input from perfusion and oxygen consumption, BOLD sequences are often, rather loosely, taken as a substitute for perfusion measurements. However, in cases where blood volume or oxygen extraction vary significantly under stimulation, this is inadequate.
More than other organs, muscle under stimulation undergoes large changes of perfusion and microvascular volume, blood and tissue oxygenation, and muscle water content. These parameters modulate the intensity of the nuclear magnetic resonance (NMR) water signal so that the interpretation of muscle water signal variations observed during a stimulation protocol is particularly difficult.
Inflow effect is assumed by Wigmore et al. (7) to be a minor contribution to signal variation. Unfortunately this assumption was not verified by the authors, although this would have been possible by applying saturation of up- and downstream planes immediately before image acquisition.
Conversely, perfusion is assumed to be a major contribution. From Fig. 4, the mean signal increase at 60% of maximal voluntary contraction is +1.8%. If this were related only to T1 shortening induced by perfusion, at temporal resolution 500 ms, this would imply a hyperhemic perfusion increase of 180 ml·100 ml–1·min–1. This does not seem plausible at the end of a 10-s isometric contraction.
Blood oxygenation variations modulate muscle signal during ischemia and exercise. Lebon et al. (2) report a signal decrease of 4% during ischemia in gastrocnemius, at 3 T, when the leg is aligned with the magnetic field, a value similar to the 4% signal decrease measured by Wigmore et al. (7) in tibialis at 1.5 T. Meyer et al. (3), in a protocol similar to Wigmore et al.'s but with shorter contractions and longer echo time, suggest that intravascular BOLD makes a major contribution to the signal variations they observed. Their interpretation is supported by the near-infrared spectroscopy measurements they performed with the same protocol on eight subjects, and by control measurements at 3 T showing signal variations two times higher than at 1.5 T, whereas perfusion-induced signal variations are independent of magnetic field intensity.
Microvascular blood volume, which may vary in such protocols, strongly modulates the muscle water signal measured with BOLD-weighted sequences.
Muscle water T2 increase after exercise is another potential contributor to signal variations.
In conclusion, the interpretation proposed by Wigmore et al. (7) suffers from a major drawback: in their experimental conditions, many factors may contribute to the signal variation and the influence of none of them can be discarded. This difficulty could be avoided by the use of arterial spin labeling (ASL) perfusion techniques that surprisingly they criticize at length in the discussion. Recently ASL measurements of human muscle perfusion were proposed with a temporal resolution more adequate for stimulation studies than that of the first publication (6). Frank et al. (1) proposed an ASL technique yielding 10-s temporal resolution, at 1.5 T. They carefully discussed the influence of relaxation times and transit times upon the perfusion values. Raynaud et al. (5) proposed an ASL technique with temporal resolution 2 s on the basis of a reasonable hypothesis of the transit time of blood in muscle and validated against plethysmography during a stress test.
Although they are technically demanding because signal variations are of the order of the percent, ASL sequences are the only way to perform absolute quantification of perfusion in vivo noninvasively and to minimize the influence of other compounding factors. We believe it is mandatory to use this methodology to fully exploit the wealth of information NMR can provide in human muscle physiology.
REFERENCES
The first point made by Professor Leroy-Willig is that our signal intensity changes are sensitive to blood oxygenation changes via the BOLD effect (2). Although this is the case, as we discuss in the paper, the short echo time (18.6 ms) will weight the signal intensity toward a T1 dependence. Pilot work indicated that when using echo times above 30 ms there is a decrease in signal intensity change, suggesting the onset of a significant T2* effect. This conclusion is consistent with those of Meyer et al. (1), who reported that postcontraction signal intensity changes observed using long-TE acquisitions reflect primarily an intravascular BOLD effect. The strong correspondence between our functional MRI (fMRI) and plethysmography data support the interpretation that our fMRI measure primarily reflects changes in blood volume.
The second point relates to inflow effects, which we consider to have minimal impact on our signal intensity changes. This conclusion was also based on the report by Meyer et al. (1), who observed that varying the repetition time did not affect the relative signal intensity change under similar, although not identical, experimental conditions. We agree that additional experiments can be designed to address this issue.
The next concern expressed by Professor Leroy-Willig relates to absolute flow changes in response to the contractions. It is not clear to us how the value of 180 ml·100 ml–1·min–1 was derived, as it is not appropriate to attempt absolute quantitation using our fMRI technique. For this reason, we reported that the increase in signal intensity after the 60% maximal voluntary contraction was approximately 49% of the maximal hyperemia obtained after cuff occlusion, which is a reasonable value. We were careful in this paper to avoid any suggestion that this technique can be used to estimate absolute flow; the discussion section deals with this limitation and suggests that, for experiments in which absolute quantitation of flow is of interest, the ASL technique is more appropriate.
We reported a 4% increase (not decrease) in signal intensity during the hyperemic period after cuff occlusion. Although this could in principle reflect a positive BOLD response brought about by reoxygenation of the muscle, we do not feel that Professor Leroy-Willig's direct comparison of our results with those of Meyer et al. (1) is appropriate, because those investigators used a TE that was significantly longer than ours and was therefore much more sensitive to oxygenation effects.
Finally, Professor Leroy-Willig expresses dissatisfaction with our discussion of the limitations of the ASL technique. Although we agree with her points—and in fact include them in our discussion—we retain the belief that ASL is the correct technique for estimating absolute levels of blood flow, but in current practice is inadequate for studies that require high temporal and, ultimately, spatial resolution. Indeed, our pilot work in the dorsiflexor muscles for this project indicated that the ASL method [using a flow-sensitive alternating inversion recovery (FAIR) sequence] resulted in insufficient signal-to-noise ratio and temporal resolution to be of use in answering the question we had posed.
We will continue to develop this approach; our plans include a number of experiments designed to more precisely define the physiological bases of the perfusion-related signal intensity changes. As with any scientific endeavour, the approach taken is dictated by the question that is the focus of investigation. We look forward to contributing, along with Dr. Leroy-Willig, to the discussion and advancements in this area.
REFERENCES
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
| HOME | HELP | FEEDBACK | SUBSCRIPTIONS | ARCHIVE | SEARCH | TABLE OF CONTENTS |
| Visit Other APS Journals Online |
