Effect of exercise intervention on thigh muscle volume and anatomical cross-sectional areas—Quantitative assessment using MRI



The objective of this study was to evaluate the location-specific magnitudes of an exercise intervention on thigh muscle volume and anatomical cross-sectional area, using MRI. Forty one untrained women participated in strength, endurance, or autogenic training for 12 weeks. Axial MR images of the thigh were acquired before and after the intervention, using a T1-weighted turbo-spin-echo sequence (10 mm sections, 0.78 mm in-plane resolution). The extensor, flexor, adductor, and sartorius muscles were segmented between the femoral neck and the rectus femoris tendon. Muscle volumes were determined, and anatomical cross-sectional areas were derived from 3D reconstructions at 10% (proximal-to-distal) intervals. With strength training, the volume of the extensors (+3.1%), flexors (+3.5%), and adductors (+3.9%) increased significantly (P < 0.05) between baseline and follow-up, and with endurance training, the volume of the extensor (+3.7%) and sartorius (+5.1%) increased significantly (P < 0.05). No relevant or statistically significant change was observed with autogenic training. The greatest standardized response means were observed for the anatomical cross-sectional area in the proximal aspect (10–30%) of the thigh and generally exceeded those for muscle volumes. The study shows that MRI can be used to monitor location-specific effects of exercise intervention on muscle cross-sectional areas, with the proximal aspect of the thigh muscles being most responsive. Magn Reson Med, 2010. © 2010 Wiley-Liss, Inc.

Assessment of human muscle morphology is useful for evaluating to what extent training, aging, immobilization, or disease affect muscle function. Muscle size/mass relates to maximum strength, but direct measures of strength depend on the motivation of the study participants during testing, are difficult to reproduce, and may not be measurable in disease.

Various in-vivo measures of muscle size/mass have been shown to correlate with muscle strength, such as muscle volume, anatomical cross-sectional area (ACSA, cross-section through the entire muscle, perpendicular to the axis of the muscle), and physiological cross-sectional area (PCSA, cross-section perpendicular to the muscle fiber direction, which is always larger than the ACSA) (1–8). Some studies have advocated muscle volume as the measure of choice for assessing muscle morphology (3, 5, 6, 8) and function (6, 8), and the current gold standard for this is three dimensional reconstruction of contiguous axial magnetic resonance (MR) images (3, 5, 6, 8–12). This process, however, requires imaging and reliable segmentation throughout a reproducible region of interest (ROI) or the entire length of the muscle, which is relatively time consuming. Therefore, several studies have focused on the ACSA from a single MRI slice, rather than muscle volume, to assess the relationship between muscle morphology and function (13, 14). Other studies have evaluated muscle volume, PCSA, and ACSA (4, 6–8), of which some recommend the use of muscle volume (6, 8), whereas others recommended ACSA (4, 7). It has been shown that muscle volume can be approximated in vivo from single-slice ACSA in the quadriceps (11) and in the triceps brachii (15). It is still unclear, which measure of muscle morphology is preferable, and if single slice ACSA is actually sufficient.

Several studies have investigated the morphological and functional response of the leg musculature to exercise intervention (5, 6, 10, 13, 16–18) or aging (12, 16). Some studies reported that muscle volume was more sensitive to changes induced by strength training than ACSA (5, 6, 10, 12), whereas others suggested that ACSA may be sufficient to monitor the response (13, 17). The reason for this discrepancy may be that the magnitude of training-induced muscle hypertrophy, measured as change in ACSA, varies with anatomical location, i.e., where along the muscle the measurement is performed (17, 19).

Therefore, the objectives of this study were

  • 1To assess whether the effect of a 3 month supervised exercise intervention (strength and endurance training) in untrained, perimenopausal, women can be monitored quantitatively by measuring thigh muscle volume with MRI;
  • 2to evaluate the response of the extensor (quadriceps), flexor (hamstring complex), adductor, and the sartorius muscles to the two exercise regimens;
  • 3to determine at which proximal-to-distal anatomical level the increase in the ACSA is most pronounced in these muscle groups;
  • 4to compare the sensitivity to change of the (proximal-to-distal) region-specific response with the response of the total muscle volume for the different muscle groups.

The latter comparison was targeted to answer the question, whether a single slice ACSA can adequately substitute the measurement of the total muscle volume in monitoring the effect of exercise intervention, when the location is optimized.



One-hundred-twenty subjects responded to an advertisement in a local newspaper and were then interviewed using a standardized telephone protocol. The following inclusion criteria were applied: women aged 45–55 years, no history of participation in organized sports, ≤1 h/week regular physical activity level, low cardio respiratory fitness, no chronic illness, no hormone-replacement therapy, amenorrhea of ≥60 days, and no surgery within the last 6 months (20). The study was approved by the Ethics Committee of the University of Salzburg and written consent was obtained from all participants before trial participation (20). Seventy women fulfilled the inclusion criteria and were randomly assigned to three groups: strength training (ST; n = 25), endurance training (ET; n = 25), and autogenic training (AT; n = 20). The latter group was used to control for the physical activity level of a non-exercising group (21, 22), as autogenic training is a relaxation based form of psycho-physiologically oriented bio-behavioral psychotherapy. After pretesting, about a third of the selected women achieved higher cardio respiratory fitness than age-normalized values based on the equation of Fitzgerald et al. (23), revealing greater than desired physical condition for the purpose of this study. These subjects were, therefore, excluded from the study. The women finally included were aged 50.8 ± 3.2 years (height 166.1 ± 7.8 cm, weight 72.8 ± 15.7 kg, BMI 26.5 ± 5.2 kg/cm2), and age and other anthropometric variables did not differ significantly between the three training groups (Kruskal-Wallis-Test). The final numbers of volunteers assigned to each group were ST = 16, ET = 19, and AT = 6, as seven of the initial 13 participants in the AT group did not complete the follow-up MRI exam because they either refused to further take part in the MRI component of the study or because they terminated participation before follow-up.


During the 12 week training period, the ST and ET groups performed exercises three times per week for 60 min, whereas the control group (AT) met once per week (20). Each ST session consisted of 5–10 min warm-up, 40–45 min training, and a 5–10 min cool-down period. Either classic hypertrophy or SuperSlow® exercises were performed, with the intensity being determined by the eight-repetition maximum test (20). The ET group completed supervised 60 min cycling sessions on an electronically braked and heart-rate-controlled cycle ergometer (20). Each cycling session consisted of 10 min warm-up, 40 min of continuous cycling, and a 10 min cool-down period. The cycling load was adjusted by a heart rate corresponding to 55–85% of the maximum determined in an incremental baseline ergometer test, and the cadence was set at 70–90 rpm for the entire session. After 3 weeks of familiarization, the ergometer workload was increased progressively. The once per week session of the AT group started and ended with relaxation exercises (Pilates and Yoga) in supine or sitting position, targeted predominantly at the upper body (pelvis, arms, and trunk), to minimize training effects on the lower extremities (20). The actual “training” consisted of six exercises based on perceptions of heaviness and warmth of the extremities, abdomen, breathing rhythm, and heartbeat, as described by Schultz and Luthe (24). The participants were instructed to maintain their habitual baseline physical activity level between the sessions. An in-depth description of the training methods and the effects on maximal isometric contraction and cardio-respiratory fitness has been published previously (20).

Muscle Morphology Analysis

MR imaging of both legs was performed in supine position with both legs extended, with a 1.5 T scanner and the integrated body coil (NT Intera, Phillips Medical Systems, Best, Netherlands). A T1-weighted turbo-spin-echo sequence (TR = 1541 msec, ET = 15 msec, ETL = 3, averages = 2, section thickness = 10 mm, in-plane resolution = 0.78 mm, acquisition time = 1.57 min) was used to acquire axial sections of both thighs, and the participants were instructed to remain relaxed within the scanner during the image acquisition to avoid motion. The knee extensors (quadriceps), knee flexors (hamstring complex), hip adductors, and the sartorius were segmented as individual entities in each slice, excluding bones, vessels, and connective tissue (Fig. 1). A ROI was determined between the distal end of the femoral neck proximally and the proximal end of the rectus femoris tendon distally (Fig. 1) because proximal and distal to the ROI, the muscles were difficult to identify. The test–retest precision for similar measurements (average of ACSA in 3 slices spaced at 25%, 50%, and 75% of the femur, with repositioning of the participant in the scanner) amounted to 2.4% for the total thigh ACSA, 1.7% for the quadriceps, 3.4% for the flexors, and 9.9% for the adductors, with the larger precision errors in the adductors likely due to the oblique orientation of the muscles in the upper thigh (25). The volume of the muscles was determined by numerical integration of segmented voxels. The ACSAs were then derived from the 3D reconstruction of the entire muscle volume (Fig. 1) at 10% intervals (proximal-to-distal, Fig. 2), and the mean ACSA was determined for each muscle by averaging all 11 reconstructed sections (0 to 100%).

Figure 1.

A, B: Color coded (extensors = magenta, flexors = green, adductors = red, sartorius = blue) 3D reconstruction of the thigh musculature within the region of interest (ROI). View from the distal end of the ROI (A) and view from medial (B). C–E: Segmented MRI slices of the thigh musculature (extensors = magenta, flexors = green, adductors = red, sartorius = blue), at the proximal end of the ROI (C), mid-thigh (D), and the distal end of the ROI (E).

Figure 2.

Axial MRI slices of the thigh covering the entire region of interest (ROI) investigated at 10% intervals from proximal to distal. The uppermost left image (0%) shows the most proximal slice of the ROI selected, where the femoral neck starts to diverge from the shaft. The bottom middle image (100%) shows the most distal slice of the ROI, where the muscle bridge between the vastus muscles of the quadriceps begins to break up to give room for the quadriceps tendon.

Statistical Analysis

Statistical analysis was performed using Excel 2003 (Microsoft Corporation, Redmond, WA). The training effect on muscle volume and ACSA was assessed by reporting the mean absolute and percent change between baseline and 12 week follow-up, the standard deviation (SD) of the change, and the standardized response mean (SRM = mean absolute change/SD), as a measure of the sensitivity to change. Differences in muscle volume between baseline and follow were tested for statistical differences using paired Student's t-tests. As four parallel tests (for four muscle groups) were applied in each group, a significance level of P < 0.0125 was required to indicate statistically significant change in a given muscle at a global significance level of P < 0.05. No adjustment for parallel testing of three training groups was conducted, however. Paired Student's t-test was also applied for the mean and proximal-to-distal ACSA, but significances are reported only in a descriptive context in comparison with the volume changes, to derive the spatial pattern of morphological adaptation; these were, therefore, not corrected for multiple comparisons.


In the ST group, the muscle volume of the extensors (+3.1%), flexors (+3.5%), and adductors (+3.9%) displayed a statistically significant increase during the 12 week training period, whereas the sartorius showed only borderline significance for change (+4.0%; P = 0.013; Table 1). For comparison, the mean ACSA (averaged cross-sectional areas) increased by 3.1%, 3.5%, 3.9% and 3.6%, respectively (not shown in the tables). In the ET group, the muscle volume of the extensors (+3.7%) and the sartorius (+5.1%) displayed a significant increase over the 12 weeks (Table 2), whereas the flexors (+0.5%) and adductors (+1.1%) showed much smaller (and nonsignificant) changes. In the AT group, there were no significant changes in muscle volume in any of the muscle groups during the training period, with a trend for a decrease in the extensors (−1.5%), flexors (−2.0%), and sartorius (−2.3%), and a nonsignificant increase in the adductors (+3.0%; Table 3).

Table 1. Strength Training Group (n = 16): Longitudinal Change of the Muscle Volume (cm3) and the Cross-Sectional Area (cm2) at Different Heights of the Thigh Musculature During the 3 Month Intervention Period
 Mean ChangeSD ChangeMean Change %SRMP
  1. CSA, cross sectional area; SRM, standardized response mean (mean change/standard deviation of change).

 CSA 10%2.331.865.481.250.001
 CSA 20%2.621.705.461.540.001
 CSA 30%2.751.615.251.710.001
 CSA 40%
 CSA 50%1.581.702.870.930.002
 CSA 60%0.971.771.810.550.045
 CSA 70%0.781.841.530.420.111
 CSA 80%0.341.640.720.210.414
 CSA 90%0.401.840.940.220.397
 CSA 10%0.300.986.970.300.246
 CSA 20%0.750.9810.770.770.008
 CSA 30%
 CSA 40%1.541.909.740.810.006
 CSA 50%1.521.947.050.780.007
 CSA 60%0.791.392.970.570.038
 CSA 70%0.371.261.280.300.253
 CSA 80%0.201.370.700.140.578
 CSA 90%0.291.831.150.160.531
 CSA 10%2.421.525.811.590.001
 CSA 20%2.371.935.491.230.001
 CSA 30%2.481.725.851.440.001
 CSA 40%1.981.705.121.160.001
 CSA 50%0.852.232.650.380.147
 CSA 60%
 CSA 70%−0.141.61−0.90−0.080.740
 CSA 80%0.070.990.700.070.790
 CSA 90%0.110.702.000.160.536
 CSA 10%−0.010.25−0.36−0.040.864
 CSA 20%
 CSA 30%0.080.332.590.240.355
 CSA 40%0.170.355.380.480.072
 CSA 50%
 CSA 60%0.200.376.350.530.049
 CSA 70%
 CSA 80%
 CSA 90%
Table 2. Endurance Training Group (n = 19): Longitudinal Change of the Muscle Volume (cm3) and the Cross-Sectional Area (cm2) at Different Heights of the Thigh Musculature During the 3 Month Intervention Period
 Mean ChangeSD ChangeMean Change %SRMP
  1. CSA, cross sectional area; SRM, standardized response mean (mean change/standard deviation of change).

 CSA 10%2.361.835.611.290.001
 CSA 20%2.691.455.721.860.001
 CSA 30%2.841.545.541.850.001
 CSA 40%2.111.433.931.480.001
 CSA 50%
 CSA 60%1.211.942.310.620.014
 CSA 70%1.322.132.680.620.015
 CSA 80%1.001.892.150.530.033
 CSA 90%0.942.082.170.450.065
 CSA 10%0.160.624.050.250.283
 CSA 20%0.501.147.980.430.076
 CSA 30%0.401.333.840.300.204
 CSA 40%0.291.381.830.210.367
 CSA 50%0.021.550.080.010.960
 CSA 60%0.041.560.170.030.903
 CSA 70%0.101.400.360.070.748
 CSA 80%0.251.30−0.85−0.190.412
 CSA 90%−0.311.38−1.14−0.230.337
 CSA 10%0.872.572.120.340.156
 CSA 20%1.512.503.620.600.017
 CSA 30%0.802.011.870.400.102
 CSA 40%0.041.920.100.020.932
 CSA 50%−0.042.57−0.11−0.010.951
 CSA 60%
 CSA 70%−0.051.88−0.34−0.030.905
 CSA 80%
 CSA 90%
 CSA 10%
 CSA 20%
 CSA 30%
 CSA 40%
 CSA 50%0.210.336.700.630.014
 CSA 60%
 CSA 70%
 CSA 80%
 CSA 90%
Table 3. Autogenic Training/Control Group (n = 6): Longitudinal Change of the Muscle Volume (cm3) and the Cross-Sectional Area (cm2) at Different Heights of the Thigh Musculature During the 3 Month Intervention Period
 Mean ChangeSD ChangeMean Change %SRMP
  1. CSA, cross sectional area; SRM, standardized response mean (mean change/standard deviation of change).

 CSA 10%−1.702.47−3.76−0.690.152
 CSA 20%−1.322.08−2.65−0.630.181
 CSA 30%−0.702.57−1.30−0.270.535
 CSA 40%−0.952.26−1.70−0.420.349
 CSA 50%−1.282.53−2.24−0.510.268
 CSA 60%−0.183.19−0.32−0.060.898
 CSA 70%0.212.730.410.080.856
 CSA 80%−0.752.50−1.54−0.300.493
 CSA 90%−0.552.12−1.22−0.260.554
 CSA 10%−0.240.32−6.12−0.740.129
 CSA 20%−0.861.36−12.05−0.640.181
 CSA 30%−0.441.97−3.69−0.220.611
 CSA 40%−1.081.39−6.58−0.780.115
 CSA 50%−0.581.84−2.56−0.320.474
 CSA 60%−0.802.94−2.90−0.270.533
 CSA 70%0.172.840.590.060.887
 CSA 80%−0.131.92−0.45−0.070.871
 CSA 90%0.321.601.150.200.647
 CSA 10%1.132.833.050.400.373
 CSA 20%0.952.632.530.360.414
 CSA 30%
 CSA 40%−0.142.41−0.41−0.060.890
 CSA 50%
 CSA 60%0.252.751.300.090.833
 CSA 70%0.181.341.340.130.754
 CSA 80%0.361.874.150.190.656
 CSA 90%0.640.8212.610.780.115
 CSA 10%0.100.353.470.280.521
 CSA 20%0.020.330.650.060.896
 CSA 30%−0.210.31−6.94−0.680.156
 CSA 40%0.140.355.130.390.379
 CSA 50%−0.020.20−0.64−0.090.829
 CSA 60%−0.070.07−2.30−1.050.049
 CSA 70%−0.130.19−4.26−0.680.158
 CSA 80%−0.050.14−1.75−0.350.430
 CSA 90%−0.060.15−2.38−0.440.332

In the ST group, the greatest SRMs for the ACSA of the extensors (1.71) and the flexors (0.89) were observed at the 30% (proximal-to-distal) level, whereas for the adductors and the sartorius the highest SRMs were seen at 10% (1.59) and 50% (0.74), respectively. In the extensors and adductors, these SRMs for the ACSA were considerably greater than those obtained for the muscle volume changes, whereas in the flexors and the sartorius, they were similar (Table 1). In the ET group, the greatest SRMs for the ACSA were found at the 20% (proximal-to-distal) level for the extensors (1.86), flexors (0.43), and adductors (0.60), whereas for the sartorius, the greatest SRM was seen at 90% (1.28). These values were substantially greater than those for the muscle volume changes (Table 2). In the AT group, the SRMs for the ACSA ranged between −1.05 and +0.78 at the various levels.


We here investigated the effect of a 3 month supervised exercise intervention (strength and endurance training vs. autogenic training) on the thigh musculature in untrained perimenopausal women using MRI. We specifically determined to what extent the different thigh muscle groups are involved in adaptation, at which proximal-to-distal anatomical level the increase in ACSA is most pronounced, and whether a single-slice measurement (at the appropriate level) provides the same (or better) sensitivity to change than measurement of the total muscle volume. We found that, with strength training (ST), the volumes of all thigh muscles increased significantly during the training period, except for the sartorius, which reached borderline significance. With endurance training (ET), only the extensors and the sartorius showed a significant increase in muscle volume, whereas the flexors and adductors showed small and inconsistent changes. With autogenic training (AT) used as a control intervention, no relevant change was observed in any muscle group. For most muscle groups (except for the sartorius), the magnitude and sensitive to change (SRM) in the proximal aspect of the thigh was greater than (or equal to) changes in muscle volume.

A limitation of this study is the low number of controls (AT), which resulted from the relatively high number of dropouts in this group. A potential reason for the relatively high drop out was that the women had signed up for participating in an exercise study but were then disappointed to be assigned to a control group and were thus less motivated to complete the study and/or participate in the MRI component. However, statistical tests were not performed between the intervention and control group but between baseline and follow-up within each group; the reason for examining a control group was to ensure that the changes observed over time were not due to drifts in the scanning equipment or to simple recruitment effects. In contrast to the intervention groups, the control group not only showed no significant changes over time but also much smaller SRMs and decreases in muscle volume for three of the four thigh muscle groups, whereas the intervention groups displayed increases in all muscles. A higher number of control subjects thus likely would not have affected the results. Another limitation is the inclusion of only female participants, as training effects on musculature can differ between sexes (10, 26). The study was specifically designed to assess the effects of the given training regimes on postmenopausal sarcopenia, and the results cannot be extended to men, to subjects of other ages, or to subjects with a higher degree of physical fitness at baseline.

A T1- (rather than T2-) weighted MRI sequence was chosen for the analysis, as it represents the current gold standard for anatomical applications and morphological muscle measurements. We, therefore, cannot exclude that potential muscle edema may have influenced the muscle volumes measured. However, physical exercise as performed in this study is not known to cause muscle edema, and no injuries or inflammatory diseases were present in the study participants. Another limitation of the study is the use of manual segmentation; this limits application of the technique to other conditions, such as senescent atrophy or myodystrophies, in which there are substantial contributions to the total volume from intramuscle fat. Several (semi-) automated image analysis techniques have been proposed for assessment of muscle morphology (fuzzy clustering algorithms (27), stereological methods (28), and thresholding (29)), or muscle composition (signal intensity assessment), which have the potential to substantially reduce the segmentation time, or exclude intramuscular fat from the measurement. However, these techniques are currently not able to reliably differentiate between thigh muscle groups, which was the focus of the current exercise intervention study. Future studies will have to show whether automated or semiautomated segmentation techniques can provide similar results.

To our knowledge, this is the first study to analyze the effect of different types of training intervention on the muscle groups of the thigh separately. The differential effect on muscle groups is something that cannot be assessed by measuring muscle strength directly, as strength tests depend on the simultaneous activation of several synergistic muscles. MRI measurement of muscle morphology is also more objective, more reproducible, and less influenced by the motivation of the participant, and can be used in situations where standard strength tests cannot easily be applied, like postoperative immobilization.

Our study confirms that MRI and digital image analysis are sensitive to exercise-induced changes in thigh muscle morphology, with the changes in quadriceps volume being in a similar range as in previous reports (6, 10, 13, 17–19, 26). Different training regimes (strength vs. endurance) affected the muscle groups of the thigh to various extents. Although with strength training, all muscle groups showed signs of hypertrophy, only the extensors and sartorius displayed a volume increase with ET. A reasonable explanation is that during cycling (ET) the flexors and adductors were not adequately stimulated, to induce an adaptation (20). Ring et al. reported a significant increase in maximal voluntary isometric contraction and an increase in the ratio of force production to leg-fat free mass in the AT group, an effect that they could not explain (20). In the here presented part of the study, the AT group showed a trend toward decreasing rather than increasing muscle volumes during the observation period and no significant changes. In a leg-specific training study, Narici et al. also observed an increase in maximal voluntary contraction of the quadriceps in the untrained leg, whereas the ACSA showed no change (17). Possibly, participation in a trial has motivational effects that affect measurements of muscle force but do not affect the measurement of muscle morphology (i.e., volume or ACSA), highlighting again the additional value of morphometric measurements.

Muscle volume is considered to represent a good estimate of muscle strength and function (6, 8), but volumetric measurements from MRI images are time consuming. Therefore, we explored whether single cross-sectional slices can be used as surrogate markers for muscle volume changes during training intervention. Several studies investigating muscle hypertrophy after training showed significant increases in ACSA at one or more anatomical levels of the quadriceps (17–19, 26). However, the results of these studies suggest that changes in muscle volume and ACSA assessments may differ from each other, depending on the location of the ACSA measurement. Several authors concluded that muscle volume (or serial measurements of ACSA) are more suitable than single mid-thigh ACSA measurements, to monitor hypertrophy in responses to strength training (6, 10, 18). Our data show that, indeed, at mid thigh (50% proximal-to-distal level), the SRM for changes in quadriceps ACSA are less than for the mean ACSA or the quadriceps volume, but that single slice ACSA is more sensitive than volume when a proximal location (20–30%) of the thigh is selected. Our results further show that, independent of the specific training regimen, flexors and extensors of the thigh display the greatest response to training at 20 to 30% distance from the femoral neck to the rectus femoris tendon, and the adductors at 10 to 20%. These findings suggest that a single-slice MRI measurement should be preferably placed in the proximal aspect of the thigh, as opposed to common mid-thigh measurements (6, 7, 11, 13), when monitoring the response of the thigh muscles to exercise.

In conclusion, this study shows that MRI can be effectively used to monitor location-specific effects of exercise intervention on muscle cross-sectional areas, with the proximal aspect of the thigh muscles being most responsive. Differential effects of strength and ET on muscle morphology can be assessed for different muscle groups of the thigh and single-slice measurements obtained at specific locations are equal or superior to measurement of muscle volumes. These findings are important in context of the design of scientific and clinical studies on muscle adaptation, where muscle morphology is measured as an (imaging) endpoint.


The authors would like to thank Christina Iuga for segmentation of the muscles in the MR images.