Muscle volume reconstruction from several short magnetic resonance imaging sequences

The gold standard to determine muscle morphological parameters is magnetic resonance imaging (MRI). To measure large muscles like the vastus lateralis (VL) in one sequence, scanners with a large field of view (FOV) and a high flux density are needed. However, large scanners are expensive and not always available. The purpose of the current study was to develop a marker-based approach to reconstruct the VL from several separate MRI sequences, acquired with a low-field MRI scanner. The VL muscle of 21 volunteers was marked at one-third and two-third of thigh length using fish oil capsules. Three consecutive MRI sequences (i.e. proximal, medial and distal part) of the thigh were captured between the markers and the muscle insertion and origin. After a manual segmentation of the VL the muscle was reconstructed using the developed approach. The muscle volume, maximal anatomical cross-sectional area and length were 715.1 ± 93.4 cm³, 34.0 ± 4.0 cm² and 34.4 ± 2.2 cm respectively. The procedure showed an average error between 0.9% and 2.2% for the reconstructed muscle volume, the averaged RMSD between the cross-sectional areas of two overlapping sequences were between 0.80 ± 0.71 cm² and 0.88 ± 0.78 cm². The proposed approach provides an appropriate accuracy for muscle volume assessment, as the estimated error for muscle volume calculation was quite small. The reconstruction quality depends mainly on the proper marker attachment and identification, as well as the spatial resolution of the image sequences. We are confident that the presented method can be used in most investigations regarding muscle morphology.