Search tips
Search criteria 


Logo of nihpaAbout Author manuscriptsSubmit a manuscriptHHS Public Access; Author Manuscript; Accepted for publication in peer reviewed journal;
J Magn Reson Imaging. Author manuscript; available in PMC 2010 August 31.
Published in final edited form as:
PMCID: PMC2930771

Calf Muscle Perfusion at Peak Exercise in Peripheral Arterial Disease: Measurement by First-Pass Contrast-Enhanced Magnetic Resonance Imaging



To develop a contrast-enhanced magnetic resonance (MR) technique to measure skeletal muscle perfusion in peripheral arterial disease (PAD).

Materials and Methods

A total of 11 patients (age = 61 ± 11 years) with moderate symptomatic PAD (ankle-brachial index [ABI] = 0.75 ± 0.08) and 22 normals were studied using an MR-compatible ergometer. PAD and normalmax (Nlmax; N = 11) exercised to exhaustion. Nllow (N = 11) exercised to the same workload achieved by PAD. At peak exercise, 0.1 mm/kg of gadolinium diethylenetriamine pentaacetic acid (Gd-DTPA) was infused at 3–4 cm3/second followed by a saline flush at the same rate. A dual-contrast gradient echo (GRE) sequence enabled simultaneous acquisition of muscle perfusion and arterial input function (AIF). The perfusion index (PI) was defined as the slope of the time-intensity curve (TIC) in muscle divided by the arterial TIC slope.


Median workload was 120 Joules in PAD, 210 Joules in Nllow, and 698 Joules in Nlmax (P < 0.001 vs. Nllow and PAD). Median PI was 0.29 in PAD (25th and 75th percentiles [%] = 0.20, 0.40), 0.48 in Nllow (25th, 75th % = 0.36, 0.62; P < 0.02 vs. PAD), and 0.69 in Nlmax (25th, 75th % = 0.5, 0.77; P < 0.001 vs. PAD). Area under the ROC-curve for PI differentiating patients from Nlmax was 0.95 (95% confidence interval [CI] = 0.77–0.99).


Peak-exercise measurement of lower limb perfusion with dual-contrast, first-pass MR distinguishes PAD from normals. This method may be useful in the study of novel therapies for PAD.

Keywords: peripheral vascular disease, perfusion, exercise, claudication

Peripheral arterial disease (PAD) is a condition characterized by flow-limiting atherosclerosis in the vessels supplying the lower limbs. Currently, between eight and 12 million Americans are affected by PAD and the incidence is expected to rise as the population ages (1). As a consequence of impaired tissue perfusion, PAD patients can experience pain, diminished exercise capacity, and tissue loss, with some ultimately requiring amputation (2). Improving blood flow is a major therapeutic goal in patients with PAD and a number of innovative approaches beyond revascularization have been investigated, although effective pharmacologic therapies that increase tissue perfusion are still lacking (35). A noninvasive technique capable of measuring tissue perfusion would be of great clinical value for assessing the severity of PAD and monitoring response to novel therapeutic interventions designed to enhance skeletal muscle perfusion.

First-pass gadolinium-enhanced MRI is a robust technique well validated for evaluating myocardial perfusion (69). First-pass gadolinium-enhanced MRI is performed using a T1-weighted sequence to visualize a gadolinium-based contrast agent in transit through tissue. Signal intensity changes in the muscle parallel contrast concentration, and time-intensity curves (TIC) can be generated in regions of interest. In myocardium, the TIC upslope correlates well with measures of micro-sphere blood flow (1012). While first-pass gadolinium-enhanced MRI has been used to evaluate blood flow in skeletal muscle among healthy individuals under predominantly nonphysiologic stress (13,14), its utility in identifying and characterizing perfusion in patients with PAD has not been extensively explored. The purpose of this study was to develop and evaluate a novel MR approach for measuring skeletal muscle perfusion semiquantitatively during exercise-induced ischemia in both mild to moderate PAD patients and normal subjects following isometric exercise within the MR environment.


Study Population

Patients between the ages of 30 and 85 years with symptoms of intermittent claudication without critical limb ischemia and an ankle-brachial index (ABI) between 0.4 and 0.9 were eligible for this study. A total of 11 patients (age = 61 ± 11 years) were recruited. Normal human subjects without risk factors for atherosclerosis were recruited from the community to serve as controls. The study was carried out using a protocol that was approved by the Human Investigation Committee at the University of Virginia Health System and all participants signed informed consent.

Study Protocol

All subjects were placed supine in the Siemens Avanto or Sonata 1.5T MR scanner (Siemens Medical Solutions, Erlangen, Germany) with the calf at the isocenter of the magnet. Monitoring of the electrocardiogram and blood pressure was performed with an InVivo 3155MVS monitor (Intermagnetics Companies, Orlando, FL, USA) throughout the study. An MRI-compatible plantar-flexion exercise ergometer (Lode, Groningen, The Netherlands) constructed for use in this protocol was then affixed to the MR table (Fig. 1). The subjects’ feet were strapped into the independently-operating pedal shoes and a rectangular circularly-polarized 20 × 50-cm2 flexible receive coil was wrapped around the calf of interest. The subjects were then placed into the magnet bore with the calf of interest at the isocenter and scout images were acquired. The maximal exercise group (normalmax [Nlmax]; N = 11, age = 46 ± 10 years) and PAD patients were then instructed to push against the pedal at a steady rate (10–12 rpm) until exhaustion or limiting symptoms developed, exercising the muscles of the foot and calf. The normal, low-level exercise group (Nllow; N = 11, age = 52 ± 6 years) was studied at a later time point and instructed to exercise to a prespecified workload, matched to that achieved by the PAD patients. Immediately following the cessation of exercise, 0.1 mm/kg of gadopentetate dimeglumine (Berlex, Montville, NJ, USA) was infused at 3–4 cm3/second through a large bore intravenous (IV) line placed in the antecubital vein followed by a 20-cm3 saline flush also infused at 3–4 cm3/second. Imaging began simultaneously, immediately following exercise cessation, and continued for a total of 100 acquisitions.

Figure 1
MR compatible, two-pedal plantar flexion ergometer (Lode, Groningen, The Netherlands) affixed to the MR table. The patient is strapped into the ergometer and placed in the magnet with their calf at the isocenter. The ergometer was designed to have a low ...

Imaging Technique

The dual-contrast sequence consisted of two interleaved slices positioned 5 cm apart with the lowest slice centered in the widest part of the calf (Fig. 2). A spoiled gradient echo (GRE) pulse sequence was modified to accommodate the simultaneous acquisition of muscle perfusion and the arterial input function (AIF). An interleaved acquisition was performed where saturation-recovery (SR) GRE (inversion time [TI] = 10 msec) was used to image the AIF and inversion-recovery (IR) GRE (TI = 320 msec) was used for muscle imaging. The delay between successive excitation pulses of the SR and IR sequences was 3.6 msec. The TI times specify the delay between the preparation pulses and the start of the data readout. The SR and IR radiofrequency pulses were nonselective. The center of k-space was measured in the first excitation pulse period (centric k-space order) and the preparation pulses were applied via the body coil. Flow compensation was used to reduce arterial flow artifact.

Figure 2
The scout image (left panel) shows the location of both the upper SR slice from which the AIF is derived and the lower IR slice, which yields the TF. Representative images from each slice are shown (right upper and lower panels) during contrast infusion ...

This pulse sequence design was used to maximize sensitivity to contrast-enhancement in the muscle while avoiding saturation of the AIF signal (Fig. 3) (15). Other parameters for both sequences included: field of view = 180 × 180, matrix size = 64 × 64, flip angle = 15°, repetition time = 900 msec, echo time = 1.86 msec, receive bandwidth = 625 Hz/pixel, and repetitions = 100. Electrocardiographic (ECG) gating was not used in order to maintain a constant repetition time, although this choice does make the method more sensitive to blood inflow effects, including pulsatile flow artifact and variations in the AIF signal. TIC were generated for regions of interest drawn in the muscle and input arteries (Fig. 4). A perfusion index (PI) was defined as the slope of the tissue TIC divided by the slope of the AIF TIC.

Figure 3
Data from phantoms plotting gadolinium concentration against signal intensity, demonstrating the extended linear range of the SR slice compared to the IR slice, as well as the improved sensitivity of the IR slice for lower concentrations of gadolinium. ...
Figure 4
Quantitative analysis of first-pass perfusion in patient #9 from Table 1. Axial SR image (left upper panel) and IR image (left lower panel) during contrast infusion at the level of the midcalf with regions of interest drawn. Note regional enhancement ...

Image Analysis

Data was analyzed using ARGUS image analysis software (Siemens). For the muscle TIC, a single operator defined at least three regions of interest approximately 1.5 cm2 in separate muscles of the calf with care taken to avoid the vasculature. To generate each TIC, the mean signal intensity was used. Because of the heterogeneous nature of blood flow to the calf depending on the muscles utilized in exercise, the muscle region with the greatest signal intensity by visual inspection during the first pass was used for the analysis. For the AIF, the saturation-recovery slices were analyzed. A smaller region of interest was used (0.3–0.4 cm2) in the vessels to avoid contamination with tissue signal and the TIC with the greatest upslope was used. Both the foot and peak of TICs were established using a linear fit model. No coil-baseline correction was used. A total of 22 randomly selected data sets were analyzed for intraobserver variability.

Statistical Analysis

Subject characteristics were compared between controls and patients using t-tests and were summarized as mean and standard deviation (SD). The Wilcoxon signed rank sum test was used to compare the centers of the PI and total work for PAD, Nlmax, and Nllow, which were then summarized as median, 25th percentile, and 75th percentile. A binormal receiver operating characteristic (ROC) curve was estimated using MedCalc statistical software. For comparison of intraobserver variability, an intraclass correlation coefficient was calculated and measurements of agreement were compared using the method of Bland and Altman (16). Linear regression analysis and all other statistical analyses were performed in SigmaStat v. 2.03 (SPSS Inc., USA).


A total of 11 patients (age = 61 ± 11 years) with mild to moderate symptomatic PAD (ABI = 0.75 ± 0.08) and 22 normal control subjects were studied. Characteristics of PAD subjects are shown in Table 1. All PAD patients and 11 normal subjects (Nlmax; age = 46 ± 10 years; P = 0.003 vs. PAD) exercised one leg to exhaustion while supine in a 1.5T MRI scanner using the custom-built, two-pedal plantar flexion ergometer. An additional 11 normal subjects (Nllow; age = 52 ± 6 years; P = not significant [NS] vs. PAD) exercised submaximally to a predefined workload similar to that achieved by the PAD subjects. Median work performed was 120 Joules in PAD (25th and 75th percentiles = 100 and 418) with a mean of 227, 210 Joules in Nllow (25th and 75th percentiles = 185 and 221; P = NS vs. PAD) with a mean of 210, and 698 Joules in Nlmax(25th and 75th percentiles = 427 and 1425; P < 0.001 vs. Nllow and PAD) with a mean of 1047 (Fig. 5).

Figure 5
Workload (Joules) achieved between the three experimental groups. The central line represents the median, the outer borders of the box represents the 25th and 75th percentiles of the data. The Nlmax group achieved a higher workload than both the PAD and ...
Table 1
Characteristics of PAD Subjects

Perfusion heterogeneity in the muscles of the lower leg was visually observed in all subjects, in part due to which muscles were used by that individual during exercise. The anterior tibialis had the greatest perfusion upslope in 67% of subjects and thus was the most frequently utilized muscle group for calculating PI. The gastrocnemius and soleus had the greatest upslope in 24% and 9% of subjects, respectively.

Median PI was 0.29 in PAD (25th and 75th percentiles = 0.20 and 0.40), 0.48 in Nllow (25th and 75th percentiles = 0.36 and 0.62; P < 0.02 vs. PAD), and 0.69 in Nlmax (25th and 75th percentiles = 0.51 and 0.77; P < 0.001 vs. PAD). PI compared between the three groups is shown in Fig. 6. Median PI was 0.57 among all control subjects (25th and 75th percentiles = 0.46 and 0.75; P < 0.001 vs. PAD). In the PAD group, the ABI failed to correlate with PI (P = NS) and neither the ABI nor the PI correlated with the workload achieved (P = NS) based on linear regression analysis. Furthermore, PI failed to correlate with workload achieved in the Nlmax group and when the entire normal cohort was pooled (P = NS). There was a weak relationship between workload and PI when data for all three groups was pooled (r = 0.379, P < 0.03).

Figure 6
PI compared between the three experimental groups. The central line represents the median, the outer borders of the box the 25th and 75th percentiles of the data. PI discriminates PAD from normals, irrespective of the workload achieved. There is a strong ...

To assess the ability of PI to discriminate disease, defined as symptomatic PAD with an ABI < 0.9 from no disease, ROC curves were generated. When both normal groups (Nlmax and Nllow) and PAD are used to generate the ROC and using a cutoff value for PI of 0.34, the area under the curve is 0.88 (95% confidence interval [CI] = 0.71–0.96). The area under the curve is 0.95 (95% CI = 0.77–0.99) when only PAD and Nlmax are analyzed together (Fig. 7). Intraobserver variability was evaluated in 22 data sets. Intraclass correlation coefficient was R = 0.910 for intraobserver variability. The Bland Altman plot is shown in Fig. 8.

Figure 7
ROC curves. The upper solid line represents the discriminatory power of PI for distinguishing PAD from Nlmax. The area under the curve is 0.95 (95% CI = 0.77–0.99). The lower dashed line represents the discriminatory power of PI for distinguishing ...
Figure 8
Bland-Altman plot of intraobserver variability. Mean bias is 0.01 with 2SD = 0.24 and −0.25. Agreement between measurements was good.


This study demonstrates that peak exercise measurement of lower limb perfusion with first-pass gadolinium-enhanced MR can distinguish mild to moderate PAD from normal controls. Employing a dual-contrast technique, we found that perfusion differences were greatest between PAD and normals who achieve their maximal workload potential (Nlmax). However, tissue perfusion remains a discriminating parameter even when workload is similar between PAD and normals. An additional advantage of this strategy is that it is semi-quantitative and does not require extensive post-processing.

A variety of noninvasive techniques have been employed for the direct assessment of human skeletal muscle blood flow including Doppler ultrasound (3,17,18), scintigraphy (19,20), and plethysmography (2123). The most widely accepted clinical standard for noninvasively measuring limb perfusion is venous occlusion strain gauge plethysmography. Aside from its inability to provide spatial localization of flow, plethysmography may underestimate perfusion and can be sensitive to changes in both body position and temperature (24). The temporally- and spatially-resolved data acquired with MR make it ideal for the study of lower limb perfusion. Consequently, a number of MR-based techniques have recently been evaluated in the assessment of lower limb blood flow (13,2528). In each of these MR studies, predominantly healthy volunteers were evaluated during reactive hyperemia and test characteristics were not assessed under low-flow conditions like that anticipated in patients with significant arterial insufficiency. Of the published techniques, those employing arterial spin labeling (25,29) and gadolinium first-pass enhancement appear most promising (13,14).

We modified a GRE pulse sequence to accommodate the simultaneous acquisition of muscle perfusion and the AIF. This approach was employed to maximize sensitivity to contrast-enhancement in the muscle while avoiding saturation of the AIF signal. An MR-compatible, custom-built, two-pedal plantar flexion ergometer enabled patients to exercise while in the scanner environment with real-time measures of work performed.

Recently, Thompson et al (14) introduced a somewhat different method for first-pass contrast-enhanced MRI in PAD. The technique relies on administration of contrast immediately following cuff inflation, both producing ischemia and allowing contrast to equilibrate in the arterial blood pool while excluding it from the lower limb. With cuff release, a true step-input of contrast was produced that coincided with hyperemic blood flow. There are three main differences between the present approach and that of Thompson et al (14): 1) the use of exercise vs. use of thigh occlusion to increase tissue blood flow; 2) the use of IR vs. SR to generate T1-contrast; and 3) the use of tissue upslope normalized by AIF upslope as a PI vs. use of a two-compartment model for estimating the unidirectional influx constant and the contrast agent distribution volume with a cuff-generated step input.

Regarding the first difference, exercise has been shown to produce higher peak flows and more accurately reproduce physiologic stress conditions (30) compared to postischemic reactive hyperemia. Furthermore, the pattern of limb perfusion at peak exercise reflects active muscle recruitment rather than differences in fiber content and capillary density that govern regional flow in reactive hyperemia models (14,31). Tolerability of prolonged cuff inflation (14) in patients with significant vascular disease in the lower limbs is also a potential concern. However, thigh occlusion may have the advantage of better reproducibility. For the second difference, IR would be expected to generate greater T1 contrast, although the pulse sequence repetition time may need to be longer. For the third difference, further studies are required to determine whether the normalized upslope or fitted model parameters are more sensitive and specific markers of PAD. For either case, an arterial input or a step function input could be used. For the thigh occlusion method, the availability of a step function input does simplify the analysis. However, if nonocclusion methods such as exercise or pharmacological vasodilation are used, then the AIF should be used to account for bolus dispersion and delay.


Aside from the small sample size, a limitation of this study was lack of validation with a separate technique capable of making absolute measures of blood flow. The PI used to compare subjects in this protocol is semi-quantitative, providing only a partial description of the perfusion system. Nevertheless, this approach accounts for variations in the AIF, which can be a source of error. In studies of myocardial first-pass perfusion, the upslope ratio tends to lead to a systematic underestimation of blood flow at higher flow rates (32). This phenomenon could have influenced our findings in the lower limbs and underestimated true differences in perfusion between groups. However, in the study by Christian et al (32), underestimation of flow was most problematic at peak flow rates greater than 200–250 mL/miniute/100 g tissue, peak values at the upper limit of that typically reported in the limbs of normal controls (14,29,33,34) and well below that described in PAD (20).

The Nlmax group was younger than the PAD group. While age-related differences between Nlmax and PAD could exaggerate differences in perfusion between these two groups, studies have demonstrated that active muscle perfusion remains intact in healthy older individuals (35,36). In all subjects in this protocol, we observed a heterogenous pattern of limb perfusion following exercise consistent with previous investigators (14,29,37). However, the pattern of perfusion heterogeneity we observed differed among some individuals, suggesting that muscle groups being utilized to perform exercise on the ergometer were inconsistent and this could have skewed our comparisons. Patients with mild to moderate PAD were studied. Patients with more severe forms of PAD with critical limb ischemia and occlusion of the superficial femoral artery (SFA) and other vessels may not be candidates for this approach as measuring the arterial input would not be feasible.

In-flow effects could result in distortions of the measured AIF curves as seen in Fig. 4. To minimize these effects, we applied flow-compensated gradient waveforms and used an echo train length of 1, whereas prior studies (13) used an echo train length of 2, which amplified flow artifact, and no flow compensation. Also, in data processing, we estimated the slope of the AIF, which is less sensitive to AIF oscillations than the de-convolution analysis used previously (13). Pulsatile arterial flow did lead to ghosting artifact and small oscillations in the measured AIFs in the present study.

Test-retest reliability studies were not performed for these measures. Variability between sites with regard to the details of the pulse sequence, receive coil, exercise protocol, and patient characteristics may limit reproducibility of these methods to other centers. If surface coils are used for signal reception, coil profile effects will need to be accounted for before calculating TIC slopes.

Potential Applications and Future Directions

Improving blood flow at a tissue level is a major therapeutic goal in the treatment of patients with PAD. While several therapies proven beneficial in the treatment of symptomatic PAD ostensibly work through enhancement of skeletal muscle perfusion (38,39), the true mechanism of action of these therapies has remained elusive because techniques capable of reliably measuring tissue blood flow have been lacking. Correlation of clinical efficacy with changes in tissue blood flow at peak exercise will help us to better understand our current treatment approach and will be critical in the development and validation of new therapeutic strategies for PAD patients. Because the skeletal muscle does not remain a passive bystander during the development of PAD, association of tissue perfusion with noninvasive measures of skeletal muscle metabolism and cellular energetics will also be important in understanding patient symptoms, performance measures, and clinical outcomes (40).

Assessment of lower limb perfusion with first-pass contrast enhancement will likely benefit from the use of higher field strength magnets. In a recent comparison of myocardial first-pass perfusion at 1.5T and 3.0T, the authors noted significant improvement in image quality with a 60% reduction in voxel size (41). Furthermore, the improved signal-to-noise ratio (SNR) at 3.0T may facilitate detection and measurement of resting limb perfusion allowing for simplified quantitation of perfusion reserve. Arterial spin labeling (ASL) is another promising technique for measuring limb perfusion that relies primarily on changes in apparent perfusion-dependent T1 relaxation times and does not require exogenous contrast agents. Hence, the method is safe and allows for repeated measurements with excellent temporal and spatial resolution. Measurement of perfusion with ASL will also improve at higher fields because of the increases in T1 relaxation times.

In conclusion, peak exercise measurement of skeletal muscle perfusion with a dual-contrast, first-pass contrast-enhanced technique is technically feasible and distinguishes PAD patients from matched controls at similar workloads. Differences in lower limb perfusion are most pronounced between PAD patients and normals who achieve their maximal workload capacity. The semiquantitative PI is relatively simple to acquire and analyze. While impractical for routine screening for PAD, this approach to measurement of tissue perfusion may aid in the assessment of disease severity and could prove valuable in the evaluation and development of novel therapies


Contract grant sponsor: National Institute of Health (NIH), National Heart, Lung, Blood Institute (NHLBI); Contract grant number: RO1 HL075792.


Walter J. Rogers is deceased.


1. American Heart Association. Heart disease and stroke statistics, 2003. Dallas, TX: American Heart Association; 2003.
2. Weitz JI, Byrne J, Clagett GP, et al. Diagnosis and treatment of chronic arterial insufficiency of the lower extremities: a critical review. Circulation. 1996;94:3026–3049. [PubMed]
3. Lazarous DF, Unger EF, Epstein SE, et al. Basic fibroblast growth factor in patients with intermittent claudication: results of a phase I trial. J Am Coll Cardiol. 2000;36:1239–1244. [PubMed]
4. Rajagopalan S, Mohler ER, III, Lederman RJ, et al. Regional angiogenesis with vascular endothelial growth factor in peripheral arterial disease: a phase II randomized, double-blind, controlled study of adenoviral delivery of vascular endothelial growth factor 121 in patients with disabling intermittent claudication. Circulation. 2003;108:1933–1938. [PubMed]
5. Mohler ER, Rajagopalan S, Olin J, Trachtenberg JD, Pak R, Crystal RG. Adenoviral-mediated gene transfer of vascular endothelial growth factor in critical limb ischemia: safety results from a phase I trial. Vasc Med. 2003;8:9–13. [PubMed]
6. Al Saadi N, Nagel E, Gross M, et al. Noninvasive detection of myocardial ischemia from perfusion reserve based on cardiovascular magnetic resonance. Circulation. 2000;101:1379–1383. [PubMed]
7. Schwitter J, Nanz D, Kneifel S, et al. Assessment of myocardial perfusion in coronary artery disease by magnetic resonance: a comparison with positron emission tomography and coronary angiography. Circulation. 2001;103:2230–2235. [PubMed]
8. Nagel E, Klein C, Paetsch I, et al. Magnetic resonance perfusion measurements for the noninvasive detection of coronary artery disease. Circulation. 2003;108:432–437. [PubMed]
9. Paetsch I, Jahnke C, Wahl A, et al. Comparison of dobutamine stress magnetic resonance, adenosine stress magnetic resonance, and adenosine stress magnetic resonance perfusion. Circulation. 2004;110:835–842. [PubMed]
10. Wilke N, Simm C, Zhang J, et al. Contrast-enhanced first pass myocardial perfusion imaging: correlation between myocardial blood flow in dogs at rest and during hyperemia. Magn Reson Med. 1993;29:485–497. [PubMed]
11. Epstein FH, London JF, Peters DC, et al. Multislice first-pass cardiac perfusion MRI: validation in a model of myocardial infarction. Magn Reson Med. 2002;47:482–491. [PubMed]
12. Wilke N, Jerosch-Herold M, Wang Y, et al. Myocardial perfusion reserve: assessment with multisection, quantitative, first-pass MR imaging. Radiology. 1997;204:373–384. [PubMed]
13. Lutz AM, Weishaupt D, Amann-Vesti BR, et al. Assessment of skeletal muscle perfusion by contrast medium first-pass magnetic resonance imaging: technical feasibility and preliminary experience in healthy volunteers. J Magn Reson Imaging. 2004;20:111–121. [PubMed]
14. Thompson RB, Aviles RJ, Faranesh AZ, et al. Measurement of skeletal muscle perfusion during postischemic reactive hyperemia using contrast-enhanced MRI with a step-input function. Magn Reson Med. 2005;54:289–298. [PMC free article] [PubMed]
15. Gatehouse PD, Elkington AG, Ablitt NA, Yang GZ, Pennell DJ, Firmin DN. Accurate assessment of the arterial input function during high-dose myocardial perfusion cardiovascular magnetic resonance. J Magn Reson Imaging. 2004;20:39–45. [PubMed]
16. Bland JM, Altman DG. Statistical methods for assessing agreement between two methods of clinical measurement. Lancet. 1986;1:307–310. [PubMed]
17. Radegran G, Saltin B. Muscle blood flow at onset of dynamic exercise in humans. Am J Physiol. 1998;274:H314–H322. [PubMed]
18. Whitney RJ. The measurement of volume changes in human limbs. J Physiol. 1953;121:1–27. [PubMed]
19. Ament W, Lubbers J, Rakhorst G, et al. Skeletal muscle perfusion measured by positron emission tomography during exercise. Pflugers Arch. 1998;436:653–658. [PubMed]
20. Depairon M, Zicot M. The quantitation of blood flow/metabolism coupling at rest and after exercise in peripheral arterial insufficiency, using PET and 15-0 labeled tracers. Angiology. 1996;47:991–999. [PubMed]
21. Sandberg M, Zhang Q, Styf J, Gerdle B, Lindberg LG. Non-invasive monitoring of muscle blood perfusion by photoplethysmography: evaluation of a new application. Acta Physiol Scand. 2005;183:335–343. [PubMed]
22. Duet M, Virally M, Bailliart O, et al. Whole-body (201)Tl scintigraphy can detect exercise lower limb perfusion abnormalities in asymptomatic diabetic patients with normal Doppler pressure indices. Nucl Med Commun. 2001;22:949–954. [PubMed]
23. Kalliokoski KK, Laaksonen MS, Takala TO, Knuuti J, Nuutila P. Muscle oxygen extraction and perfusion heterogeneity during continuous and intermittent static exercise. J Appl Physiol. 2003;94:953–958. [PubMed]
24. Jorfeldt L, Vedung T, Forsstrom E, Henriksson J. Influence of leg position and environmental temperature on segmental volume expansion during venous occlusion plethysmography. Clin Sci (Lond) 2003;104:599–605. [PubMed]
25. Raynaud JS, Duteil S, Vaughan JT, et al. Determination of skeletal muscle perfusion using arterial spin labeling NMRI: validation by comparison with venous occlusion plethysmography. Magn Reson Med. 2001;46:305–311. [PubMed]
26. Richardson RS, Haseler LJ, Nygren AT, Bluml S, Frank LR. Local perfusion and metabolic demand during exercise: a noninvasive MRI method of assessment. J Appl Physiol. 2001;91:1845–1853. [PubMed]
27. Toussaint JF, Kwong KK, Mkparu FO, Weisskoff RM, LaRaia PJ, Kantor HL. Perfusion changes in human skeletal muscle during reactive hyperemia measured by echo-planar imaging. Magn Reson Med. 1996;35:62–69. [PubMed]
28. Lebon V, Carlier PG, Brillault-Salvat C, Leroy-Willig A. Simultaneous measurement of perfusion and oxygenation changes using a multiple gradient-echo sequence: application to human muscle study. Magn Reson Imaging. 1998;16:721–729. [PubMed]
29. Frank LR, Wong EC, Haseler LJ, Buxton RB. Dynamic imaging of perfusion in human skeletal muscle during exercise with arterial spin labeling. Magn Reson Med. 1999;42:258–267. [PubMed]
30. Andersen P, Saltin B. Maximal perfusion of skeletal muscle in man. J Physiol. 1985;366:233–249. [PubMed]
31. Gray SD, Renkin EM. Microvascular supply in relation to fiber metabolic type in mixed skeletal muscles on rabbits. Microvasc Res. 1978;16:406–425. [PubMed]
32. Christian TF, Rettmann DW, Aletras AH, et al. Absolute myocardial perfusion in canines measured by using dual-bolus first-pass MR imaging. Radiology. 2004;232:677–684. [PubMed]
33. Saltin B. Capacity of blood flow delivery to exercising skeletal muscle in humans. Am J Cardiol. 1988;62:30E–35E. [PubMed]
34. Andersen P, Saltin B. Maximal perfusion of skeletal muscle in man. J Physiol. 1985;366:233–249. [PubMed]
35. Proctor DN, Newcomer SC, Koch DW, Le KU, MacLean DA, Leuenberger UA. Leg blood flow during submaximal cycle ergometry is not reduced in healthy older normally active men. J Appl Physiol. 2003;94:1859–1869. [PubMed]
36. Martin WH, III, Ogawa T, Kohrt WM, et al. Effects of aging, gender, and physical training on peripheral vascular function. Circulation. 1991;84:654–664. [PubMed]
37. Laughlin MH. Distribution of skeletal muscle blood flow during locomotory exercise. In: Gonzalez NC, Roger Fedde M, editors. Oxygen transfer from atmosphere to tissues. New York: Plenum; 1988. pp. 87–102.
38. Regensteiner JG, Hiatt WR. Current medical therapies for patients with peripheral arterial disease: a critical review. Am J Med. 2002;112:49–57. [PubMed]
39. Mohler ER, III, Hiatt WR, Creager MA. Cholesterol reduction with atorvastatin improves walking distance in patients with peripheral arterial disease. Circulation. 2003;108:1481–1486. [PubMed]
40. Isbell DC, Berr SS, Toledano AY, et al. Delayed calf muscle phosphocreatine recovery after exercise identifies peripheral arterial disease. J Am Coll Cardiol. 2006;47:2289–2297. [PMC free article] [PubMed]
41. Strach KA, Meyer C, Naehle CP. High resolution myocardial perfusion imaging at 3.0T: comparison to standard 1.5T perfusion studies and diagnostic accuracy in patients with suspected CAD [Abstract] J Cardiovasc Magn Reson. 2006;8:7–8.