Determination of Mouse Skeletal Muscle Architecture Using Three-Dimensional Diffusion Tensor Imaging Anneriet M. Heemskerk, 1 * Gustav J. Strijkers, 1 Anna Vilanova, 2 Maarten R. Drost, 3 and Klaas Nicolay 1 Muscle architecture is the main determinant of the mechanical behavior of skeletal muscles. This study explored the feasibility of diffusion tensor imaging (DTI) and fiber tracking to noninva- sively determine the in vivo three-dimensional (3D) architecture of skeletal muscle in mouse hind leg. In six mice, the hindlimb was imaged with a diffusion-weighted (DW) 3D fast spin-echo (FSE) sequence followed by the acquisition of an exercise- induced, T 2 -enhanced data set. The data showed the expected fiber organization, from which the physiological cross-sectional area (PCSA), fiber length, and pennation angle for the tibialis anterior (TA) were obtained. The values of these parameters ranged from 5.4 –9.1 mm 2 , 5.8 –7.8 mm, and 21–24°, respectively, which is in agreement with values obtained previously with the use of invasive methods. This study shows that 3D DT acquisi- tion and fiber tracking is feasible for the skeletal muscle of mice, and thus enables the quantitative determination of mus- cle architecture. Magn Reson Med 53:1333–1340, 2005. © 2005 Wiley-Liss, Inc. Key words: muscle fiber architecture; diffusion tensor imaging; fiber tracking; exercise-induced T 2 enhancement; mice Muscle architecture is the main determinant of the me- chanical behavior of skeletal muscles (1–3). This architec- ture, which is defined as the arrangement of muscle fibers relative to the axis of force generation (3), is characterized by various parameters, including muscle length, fiber length, pennation angle, and physiological cross-sectional area (PCSA). The pennation angle is the angle between the muscle fibers and the tendon plate, while the PCSA is the sum of the cross-sectional areas of all fibers. The PCSA is the architectural parameter that is directly proportional to the maximum force generated by the muscle (3). However, the PCSA is also the parameter that is the most difficult to measure, and is therefore normally indirectly determined from the muscle volume and fiber length (4). The fiber length and pennation angle can be determined by ultrasonography (US) (5) and anatomical reconstruction methods. However, both of these techniques have severe limitations. US does not account for spatial variation in fiber length and orientation within a muscle. The disad- vantage of traditional anatomical reconstruction tech- niques is that they are restricted to ex vivo preparations (2), which precludes the use of longitudinal studies. An alternative, noninvasive method used to measure the skel- etal muscle fiber structure is magnetic resonance imaging (MRI). This method may enable the reconstruction of whole muscles and provide the desired architectural pa- rameters, including the PCSA. MRI offers researchers an opportunity to study the mi- crostructure of tissues such as brain white matter (6,7), heart muscle (8,9), and skeletal muscle (6,10 –14) using diffusion tensor imaging (DTI). DTI relies on the fact that the self-diffusion of water in tissue is restricted by mem- branes and other cellular constituents, resulting in an ap- parent diffusion coefficient (ADC), which is lower than the free diffusion coefficient and is orientation-dependent for elongated structures. One can characterize the orientation dependency by measuring the diffusion with MRI in at least six directions, and the DT can be calculated from this measurement. The eigenvalues and eigenvectors derived from this tensor provide information on the local tissue geometry. It has been shown that for skeletal muscle the eigenvector with the largest eigenvalue (i.e., the principal eigenvector) corresponds to the long axis of the muscle fibers (11,12). These local, voxel-based directions can be combined by a fiber-tracking algorithm (15,16) to recon- struct the paths of muscle fibers in the tissue (13). The mechanisms of force loss that occur during progres- sive muscular diseases caused by genetic defects (e.g., Duchenne’s muscular dystrophy and Pompe’s disease) are increasingly being studied in genetically modified mice (17,18). Such diseases generally cause both progressive muscle wasting and a decrease in maximal tension. This wasting also causes changes in muscle architecture. As stated above, the fiber architecture is the main determinant of mechanical behavior. Therefore, to determine any ac- tual loss in muscle quality, one must know the architec- ture. The DTI-based fiber-reconstruction method is ex- pected to facilitate the study of architectural changes in mice, which is a most challenging application because of the small size of mouse muscles. The aim of this study was to develop the methods to determine the in vivo 3D architecture of mouse skeletal muscle, using DTI. We will show that with this approach it is possible to calculate the PCSA and determine the fiber length and the pennation angle of the tibialis anterior (TA) muscle, which can be easily identified and has been ex- tensively studied. The methods used were an in vivo 3D acquisition, a fiber-tracking algorithm, and analysis and visualization tools. For part of this study we focused on the possibility that fibers might erroneously cross muscle borders. We indirectly delineated these borders using T 2 1 Biomedical NMR, Department of Biomedical Engineering, Eindhoven Univer- sity of Technology, Eindhoven, The Netherlands. 2 Biomedical Image Analysis, Department of Biomedical Engineering, Eind- hoven University of Technology, Eindhoven, The Netherlands. 3 Department of Movement Sciences, Maastricht University, Maastricht, The Netherlands. *Correspondence to: Anneriet Heemskerk, Department of Biomedical Engi- neering, Eindhoven University of Technology, P.O. Box 513, 5600 MB, Eind- hoven, The Netherlands. E-mail: a.m.heemskerk@tue.nl Received 12 October 2004; revised 15 December 2004; accepted 29 Decem- ber 2004. DOI 10.1002/mrm.20476 Published online in Wiley InterScience (www.interscience.wiley.com). Magnetic Resonance in Medicine 53:1333–1340 (2005) © 2005 Wiley-Liss, Inc. 1333