In Vivo Diffusion Tensor Imaging and Tractography of Human Thigh Muscles in Healthy Subjects

Transcription

1 Musculoskeletal Imaging Original Research Kermarrec et al. Diffusion Tensor Imaging and Tractography of Thigh Muscles Musculoskeletal Imaging Original Research Erwan Kermarrec 1 2 Jean-François udzik 1 Chadi Khalil 1 Vianney Le Thuc 1,2,3 Caroline Hancart-Destee 1 nne Cotten 1 Kermarrec E, udzik JF, Khalil C, Le Thuc V, Hancart-Destee C, Cotten Keywords: diffusion tensor imaging, fiber tracking, MRI, muscle, thigh DOI: /JR Received July 25, 2009; accepted after revision pril 16, Service de Radiologie et Imagerie Musculosquelettique, Centre de Consultation et Imagerie de l ppareil Locomoteur, CHRU Lille, oulevard du Pr. J. Leclercq, Lille, France. ddress correspondence to E. Kermarrec (erwankermarrec@yahoo.fr). 2 Université Lille Nord de France, Lille, France. 3 Institut National de la Santé et de la Recherche Médicale, Loos, France. WE This is a Web exclusive article. JR 2010; 195:W352 W X/10/1955 W352 merican Roentgen Ray Society In Vivo Diffusion Tensor Imaging and Tractography of Human Thigh Muscles in Healthy Subjects OJECTIVE. The aims of this study were to assess whether similar measurements of mean apparent diffusion coefficient and fractional anisotropy in muscles can be obtained with regions of interest drawn on cross-sectional diffusion tensor images and tractography and to assess whether water diffusivity in human thigh muscles is influenced by muscular compartment, age, and sex. SUJECTS ND METHODS. Sixteen healthy volunteers (eight women, eight men) participated in this study. The right thigh of each subject was imaged with diffusion tensor imaging, and the mean apparent diffusion coefficient and fractional anisotropy values were calculated for each muscle of the quadriceps femoris and hamstrings. Fiber tracking was performed with a line propagation technique from the regions of interest drawn on cross-sectional diffusion tensor images. RESULTS. The water diffusivity parameters obtained with tractography did not differ significantly from those obtained with diffusion tensor imaging in the three regions of interest evaluated in each muscle. The mean apparent diffusion coefficient of women ( mm 2 /s) was similar to that of men ( mm 2 /s). Women and men had identical fractional anisotropy values (0.26). The fractional anisotropy value in young volunteers (0.27) was similar to that in older subjects (0.26). The hamstrings had a lower mean apparent diffusion coefficient ( mm 2 /s) than the quadriceps femoris ( mm 2 /s), but the quadriceps femoris had a significantly lower fractional anisotropy value (0.25) than the hamstrings (0.28). CONCLUSION. Our study showed that the water diffusivity values (mean apparent diffusion coefficient and fractional anisotropy) of the thigh muscles did not differ significantly with respect to sex or age of the subject. The quadriceps femoris and the hamstrings did have different mean apparent diffusion coefficient and fractional anisotropy values, which may reflect differences in hydration and muscular architecture. S triated muscle is highly ordered tissue characterized by an anisotropic arrangement of fibers. The molecular motion of water preferentially occurs along the long axis of the muscle fibers, whereas it is reduced in the direction perpendicular to this axis [1, 2]. The random microscopic motion of water in tissues can be measured with MRI that includes diffusion-weighted imaging (DWI) sequences [3]. In diffusion tensor imaging (DTI), measurements related to the molecular motion of water in space in anisotropic tissues are used to acquire information about the orientation and architectural organization of tissue [4]. With tractography, or fiber tracking, 3D fiber tracks are visualized by means of mathematic representation of a physical phenomenon. DWI and DTI have clinical applications in evaluation of the CNS [5], but their use in musculoskeletal imaging is still emerging. DTI assessments of the architecture of the calf muscles of humans [6 12] and animals [2, 13, 14] have been described. The feasibility of in vivo DTI and tractography of human thigh muscles also has been reported [15]. These imaging techniques are a reliable way to evaluate muscular microstructural parameters (apparent diffusion coefficient [DC] and fractional anisotropy [F]) and are a good approach to the assessment of muscle shape and the orientation of muscle fibers [15]. udzik et al. [15] found that tractography can be used to obtain the measurements needed to calculate the mean DC and F of muscles. In previous studies [6 10], those values were obtained solely by positioning one or several regions of interest (ROIs) on cross-sectional DT images. To the best of our knowledge, even though the thigh W352 JR:195, November 2010

2 Diffusion Tensor Imaging and Tractography of Thigh Muscles Fig year-old female healthy volunteer. Cross-sectional images from midpart of thigh., T1-weighted image., Fat-suppressed T2-weighted STIR image. C, Spin-echo single-shot diffusion-weighted echo-planar image. muscles are frequently affected in various conditions, including trauma and inflammatory disorders, no comparison of the mean DC and F values obtained with DTI and tractography has been performed, nor have data been published concerning the potential influence of age, sex, and muscle shape on the mean DC and F of the thigh muscles. The aim of our study was to assess whether water diffusivity in human thigh muscles is influenced by muscular compartment, age, and sex. This goal required first assessing whether the mean DC and F values in muscles depend on the measurement technique used, that is, tractography or drawing ROIs on cross-sectional DT images. Subjects and Methods Our study was approved by our institutional review board and ethics committee. Sixteen healthy volunteers (eight women, eight men; mean age, 47.1 years; range, years) were included. Each volunteer provided informed consent. The participants were grouped by age as young (mean age, < 47.1 years) and old (mean age, > 47.1 years). Each age group had eight subjects, four men and four women. The age criterion was chosen on the basis of results of a previous study [8]. No volunteer had symptoms or any disease, and to avoid changes in the hydration of the muscles, no subject participated in sports or intense activity during the 48 hours before the study imaging. MRI Protocol and Image cquisition ll MRI was performed with a 1.5-T full-body system (chieva, Philips Healthcare). The volunteers were supine feet first with their lower extremities in a relaxed state parallel to the magnetic field. The right thigh of each subject was imaged with a sensitivity-encoding body coil. The following three MRI acquisitions were successively performed (Fig. 1): T1-weighted images, fat-suppressed T2-weighted STIR images, and spin-echo single-shot echo-planar DW images. The T1- weighted and STIR images were used to confirm the normal morphologic features and signal intensity of the thigh muscles. The T1-weighted images were acquired with the following parameters: TR/TE, 542/15; total acquisition time, 5 minutes 1 second. STIR images were acquired with the following parameters: 1,500/10; inversion time, 150 milliseconds; turbo-spin echo factor, 5; total acquisition time, 6 minutes 18 seconds. These sequences were followed by a spin-echo single-shot echo-planar DWI sequence with fat-suppression. The DTI acquisition consisted of one 0 image and DT images obtained with a b factor of 400 s/ mm 2 in 32 directions. To avoid chemical shift artifacts, frequency-selective fat saturation was used to suppress the fat signal (Fig. 2). The image acquisition parameters for the diffusion images were as follows: 6,197/60; field of view, 200 mm; parallel imaging technique; sensitivity-encoding factor, 2; partial Fourier acquisition (half-acquisition factor, 0.681), recon matrix size, ; anteroposterior phase encoding; bandwidth in frequency direction, 1,367.5 Hz/pixel. Thirty-six transverse slices were acquired without spacing. The in-plane image resolution was mm with a slice thickness of 8 mm. Five independent acquisitions were performed in succession to improve signalto-noise ratio, which was in muscle. The total acquisition lasted 18 minutes 35 seconds. To match the anatomic images with the DT images, the same field of view, slice thickness, recon matrix, number of slices and slice positioning were used for the T1-weighted and fat-suppressed T2- weighted images and for the DT images. Postprocessing Diffusion registration Eddy currents and motion-related misalignment of the DT images were corrected offline with automated image registration (Philips Research Imaging Developing Environment, Philips Healthcare; IDL, ITT). ll DW images were reoriented to match the 0 images. Each series was then reoriented to match the others (affine transformation). In all cases, DT images were perfectly superimposed and aligned. The final step consisted in averaging the five reoriented acquisitions. This step made it possible to improve signal-to-noise ratio and minimize artifacts. Mean DC and F measurements The muscles studied were the quadriceps femoris (rectus femoris, vastus intermedius, vastus medialis, vastus lateralis) and the hamstrings (long head of the biceps femoris, semitendinosus, semimembranosus). The adductus muscles were excluded from this study because of difficulty in identifying the three muscular boundaries on the native DT images because these muscles have a close relation. One radiologist calculated the mean DC and F values for each muscle, first using three ROIs drawn manually on cross-sectional native DT images in the superior, middle, and inferior thirds of each muscle (Fig. 3). Each ROI yielded mean DC and F values at one level. The averages of the mean DC and F values obtained at the three levels yielded the mean DC and F values for each muscle. Three ROIs per muscle were used to limit intramuscular variability of the mean DC and F values. Fiber tracking was performed with a line propagation technique (Philips Research Imaging Developing Environment FiberTracking 6.2, Philips Healthcare) (Fig. 3). Tracking was launched from a seed ROI from which a line was propagated in both the retrograde and anterograde directions according to the main eigenvector at each voxel. C JR:195, November 2010 W353

3 Kermarrec et al. We used three ROIs per muscle (those previously drawn manually at the superior, middle, and inferior thirds of each muscle). For tractography, the following parameters were chosen in the software: F threshold, 0.12; direction threshold, The F threshold determines the F value below which voxels are excluded from the propagation algorithm. The direction threshold determines the value of the maximum angle authorized between the main eigenvectors of two adjacent voxels. The threshold values were chosen with the aim of obtaining as many fibers as possible without including aberrant fibers (Fig. 4). Once fiber tracking was performed, with a single right-click of the computer mouse on the bundle of fibers, the software generated the mean DC and F values. The mean DC and F generated with the software represented the mean DC Fig year-old male healthy volunteer., Spin-echo single-shot diffusion-weighted echo-planar image without fat suppression shows chemical shift artifacts., Fat-suppressed image shows no artifacts. Fig year-old male healthy volunteer. Example of two measurement techniques used to calculate mean apparent diffusion coefficient and fractional anisotropy values in rectus femoris., Native diffusion tensor image shows three regions of interest., Tractography image shows fiber tracking from three regions of interest. and F values measured in each voxel included in the volume. To avoid partial volume artifact, for both DTI and tractography, the ROIs chosen were slightly smaller than the cross-sectional area of each muscle. ll measurements were performed twice by the same radiologist to evaluate intraobserver reproducibility. The radiologist performed the second analysis 6 months after the first. Statistical nalysis The mean DC and F values obtained with the two measurement techniques (DTI and tractography) were compared. The mean DC and F values then were analyzed according to age, sex, and muscular compartment (quadriceps femoris versus hamstrings, i.e., anterior versus posterior). The comparisons were performed with Wilcoxon s test and the Mann-Whitney U test. The level of statistical significance was set at p < Intraobserver reproducibility was evaluated with the intraclass correlation coefficient. Intraobserver reproducibility was considered very good at R > 0.91, good at 0.71 < R < 0.91, moderate at 0.51 < R < 0.71, mediocre at 0.31 < R < 0.51, and null at R < ll statistical tests were performed with SS software (version 9.1, SS Institute). Results The T1-weighted and STIR images showed no evidence of atrophy, fatty infiltration, or edema in the thigh muscles of the healthy volunteers. The mean DC and F values of the muscles are shown in Table 1. The water diffusivity parameters obtained with tractography did not significantly differ from those obtained from three ROIs on DT images (p > 0.05, Wilcoxon s test). Likewise, the mean DC and F values in the seven muscles analyzed did not significantly differ according to sex or age (p > 0.05). The mean DC in women ( mm 2 /s) was similar to that of men ( mm 2 /s), and the F values in women and men were identical (0.26). The F value in young volunteers (0.27) was similar to that in older subjects (0.26). Even though the mean DC was lower in older subjects ( mm 2 /s versus mm 2 /s in younger volunteers), the difference was not significant (p > 0.05). The water diffusivity values in the anterior muscular compartment, however, were significantly different from those in the posterior compartment. The hamstrings had a lower mean DC ( mm 2 /s) than the quadriceps femoris ( mm 2 /s) (p = 0.001), and the quadriceps femoris had a significantly lower F value (0.25) than the hamstrings (0.28) (p = ). For all the studied thigh muscles, intraobserver reproducibility was very good or good for three-roi DTI with respect to both F and DC (0.79 < R < 0.97). Intraobserver reproducibility also was good for tractography for both F and DC values in the rectus femoris, vastus intermedius, and biceps femoris (0.71 < R < 0.89) but was imperfect for both F and DC values in the vastus lateralis and semimembranosus (R < 0.52). Discussion efore any comparison of mean DC and F values according to age, sex, and muscular compartment, our study showed that either tractography or DTI can be used to obtain mean DC and F values in normal human thigh muscles. Similar values were obtained with the two measurement techniques. W354 JR:195, November 2010

4 Diffusion Tensor Imaging and Tractography of Thigh Muscles To the best of our knowledge, this study is the first to compare the mean DC and F values of muscles obtained with these two techniques. It is important to keep in mind, however, that intraobserver reproducibility was better for DTI with three ROIs per muscle than it was for tractography. We have no clear explanation for the lack of reproducibility of both the mean DC and F values obtained with tractography of the vastus lateralis and semimembranosus. The DC values obtained in this study differ from those obtained in a previous study [15], despite use of the same imaging protocol. This discrepancy can be explained by the use of different versions of the software because the method of calculating DC was changed between FiberTracking 4.1 beta 4 and FiberTracking 6.2. This difference explains why the previous measurements were lower than those previously reported for calf muscles [7, 9, 10]. The DC measurements obtained in this study are similar or slightly greater than those obtained by Sinha et al. [9, 10] and Galban et al. [7]. C We found a significant difference between the mean DC values of the hamstrings and quadriceps femoris muscles. It is now well Fig year-old female healthy volunteer. Tractographic images obtained with different values of direction threshold in semitendinosus muscle. When deflection angle is set at higher angle, more fibers are tracked. Risk is tracking aberrant fibers., Direction threshold, 2.7., Direction threshold, C, Direction threshold, 18. D, Direction threshold, known from studies of neurologic disorders [16, 17] that an increased amount of extracellular water (vasogenic edema) is among the main causes of increased mean DC values. Therefore, the higher mean DC of the quadriceps femoris in contrast to that of the hamstrings may suggest greater hydration of the quadriceps femoris than of the hamstrings. Interestingly, thigh injuries, particularly strain, affect the hamstrings more frequently than the quadriceps femoris [18]. Several risk factors, such as eccentric exercise, two-joint muscles, a high proportion of type 2 fibers, and unipennate or bipennate structure of the muscles, may explain this difference in the distribution of thigh injuries. Lesser hydration of the hamstrings also may be a risk factor accounting for more frequent injury to those muscles. In our study, the mean DC values measured in the thigh muscles did not vary significantly by sex or age of the volunteers, although these values appeared to be lower in older subjects. These results differ from those on human calf muscles reported by Galban et al. [7], who in a study with 24 volunteers TLE 1: Mean pparent Diffusion Coefficient and Fractional nisotropy Values Muscle D Mean pparent Diffusion Coefficient ( 10 3 mm 2 /s) Diffusion Tensor Imaging Tractography Fractional nisotropy Diffusion Tensor Imaging Tractography Vastus medialis Vastus lateralis Vastus intermedius Rectus femoris Quadriceps femoris a Long head of biceps femoris Semitendinosus Semimembranosus Hamstrings b Entire thigh c a Values for vastus medialis, vastus lateralis, vastus intermedius, and rectus femoris divided by four. b Values for long head of biceps femoris, semitendinosus, and semimembranosus divided by three. c Values for all seven muscles divided by seven. JR:195, November 2010 W355

5 Kermarrec et al. found higher mean DC values in women than in men. We have no clear explanation for this discrepancy, and we do not know whether it can be explained by the fact that different muscles are involved. Moreover, to the best of our knowledge, no report in the literature shows that the hydration status of the muscles, in contrast to total body water, differs according to sex and age [19]. Further studies are needed to improve our knowledge in this area. It would be especially interesting to assess whether our values, obtained at rest, would be influenced by exercise. We found that the F value in the quadriceps femoris was significantly lower than that in the hamstrings. y definition, a tissue whose fibers evidence multiple orientations (imperfectly oriented structure) has a lower degree of anisotropy and consequently a lower F value [16]. One possible hypothesis is that the orientation of the fibers of thick muscles, especially the vastus medialis and vastus lateralis, is more variable than that of spindleshaped muscles such as the hamstrings. nother hypothesis is that the pennate structure of the muscles and consequently the geometric arrangement of the fibers within the muscle may influence F value. Sinha et al. [10] found lower F values in the soleus muscle than in the medial gastrocnemius, lateral gastrocnemius, and tibialis anterior muscles. Those authors hypothesized that the bipennate structure of the anterior part of the soleus may explain the lower F value. In our study, F values did not vary significantly with sex or age. To the best of our knowledge, the possible variation in F according to those parameters has not been studied in thigh muscles. If similar muscular organization and consequently a similar degree of anisotropy can be expected in women and men, physiologic age-related muscle atrophy may modify F in older patients. Galban et al. [8] reported lower F in only one calf muscle in 38 volunteers. Further studies with a larger sample and with a larger difference in mean age between the two groups should be performed to confirm the absence of F changes with age in thigh muscles. We acknowledge several limitations in our study. First, tractography was performed by only one radiologist, so interobserver reproducibility was not assessed. To the best of our knowledge, such an assessment has not been previously reported for DTI and tractography of muscles. The validity of these imaging techniques will require assessment of interobserver reproducibility in further studies. Second, we examined only a small number of volunteers; further studies are needed to confirm these F and mean DC values in normal thigh muscles. Our results show that similar values of mean DC and F can be obtained in normal human thigh muscles with tractography and DTI but that DTI has better intraobserver reproducibility. Our study also showed that mean DC and F values in thigh muscles did not vary significantly by sex or age. In contrast, the quadriceps femoris and hamstrings had different mean DC and F values, which may reflect differences in hydration and muscular architecture. Further studies should be undertaken to confirm these preliminary results and to evaluate applications to muscular conditions that frequently affect the thighs, such as inflammatory myopathy and injuries. References 1. Heemskerk M, Strijkers GJ, Vilanova, Drost MR, Nicolay K. Determination of mouse skeletal muscle architecture using three-dimensional diffusion tensor imaging. Magn Reson Med 2005; 53: van Doorn, ovendeerd PH, Nicolay K, Drost MR, Janssen JD. Determination of muscle fibre orientation using diffusion-weighted MRI. Eur J Morphol 1996; 34: ammer R. asic principles of diffusion-weighted imaging. Eur J Radiol 2003; 45: asser PJ, Jones DK. Diffusion-tensor MRI: theory, experimental design and data analysis a technical review. NMR iomed 2002; 15: Yu CS, Li KC, Xuan Y, Ji XM, Qin W. Diffusion tensor tractography in patients with cerebral tumors: a helpful technique for neurosurgical planning and postoperative assessment. Eur J Radiol 2005; 56: Galban CJ, Maderwald S, Uffmann K, de Greiff, Ladd ME. Diffusive sensitivity to muscle architecture: a magnetic resonance diffusion tensor imaging study of the human calf. Eur J ppl Physiol 2004; 93: Galban CJ, Maderwald S, Uffmann K, Ladd ME. diffusion tensor imaging analysis of gender differences in water diffusivity within human skeletal muscle. NMR iomed 2005; 18: Galban CJ, Maderwald S, Stock F, Ladd ME. ge-related changes in skeletal muscle as detected by diffusion tensor magnetic resonance imaging. J Gerontol 2007; 62: Sinha U, Yao L. In vivo diffusion tensor imaging of human calf muscle. J Magn Reson Imaging 2002; 15: Sinha S, Sinha U, Edgerton VR. In vivo diffusion tensor imaging of the human calf muscle. J Magn Reson Imaging 2006; 24: Lansdown D, Ding Z, Wadington M, Hornberger JL, Damon M. Quantitative diffusion tensor MRI-based fiber tracking of human skeletal muscle. J ppl Physiol 2007; 103: Zaraiskaya T, Kumbhare D, Noseworthy MD. Diffusion tensor imaging in evaluation of human skeletal muscle injury. J Magn Reson Imaging 2006; 24: Damon M, Ding Z, nderson W, Freyer S, Gore JC. Validation of diffusion tensor MRI-based muscle fiber tracking. Magn Reson Med 2002; 48: Saotome T, Sekino M, Eto F, Ueno S. Evaluation of diffusional anisotropy and microscopic structure in skeletal muscles using magnetic resonance. Magn Reson Imaging 2006; 24: udzik JF, Le Thuc V, Demondion X, Morel M, Chechin D, Cotten. In vivo MR tractography of thigh muscles using diffusion imaging: initial results. Eur Radiol 2007; 17: Le ihan D, Mangin JF, Poupon C, et al. Diffusion tensor imaging: concepts and applications. J Magn Reson Imaging 2001; 13: Gass, Niendorf T, Hirsch JG. cute and chronic changes of the apparent diffusion coefficient in neurological disorders: biophysical mechanisms and possible underlying histopathology. J Neurol Sci 2001; 186[suppl 1]:S15 S Waldén M, Hägglund M, Ekstrand J. UEF Champions League study: a prospective study of injuries in professional football during the season. r J Sports Med 2005; 39: Chumlea WC, Guo SS, Zeller CM, Reo NV, Siervogel RM. Total body water data for white adults 18 to 64 years of age: the Fels Longitudinal Study. Kidney Int 1999; 56: W356 JR:195, November 2010