Validation of 3D spino-pelvic muscle reconstructions based on dedicated MRI sequences for fat-water quantification

Abstract

Objectives: To evaluate a protocol, including MRI acquisition with dedicated sequences for fat-water quantification and semi-automatic segmentation, for 3D geometry measurement and fat infiltration of key muscles of the spino-pelvic complex. Materials and Methods: MRI Protocol: Two axial acquisitions from the thoraco-lumbar region to the patella were obtained: one T1 weighted and one based on the Dixon method, permitted to evaluate the proportion of fat inside each muscle. Muscle Reconstruction: With Muscl’X software, 3D reconstructions of 18 muscles or groups of muscles were obtained identifying their contours on a limited number of axial images (DPSO Method); 3D references were obtained only on T1 acquisitions identifying the contour of the muscles on all axial images. Evaluation: For two volunteers, three operators completed reconstructions three times across three sessions. Each reconstruction was projected on the reference to calculate the ‘point to surface’ error. Mean and maximal axial section, muscle volume, and muscle length calculated from the reconstructions were compared to reference values, and intra- and inter-operator variability for those parameters were evaluated. Results: 2xRMS ‘point to surface’ error was below 3 mm, on average. The agreement between the two methods was variable between muscles [-4.50; 8.00 %] for the mean axial section, the length and the volume. Intra- and inter-operator variability were less than 5% and comparison of variability for the Fat and T1 reconstructions did not reveal any significant differences. Discussion: Excellent inter- and intra-operator reliability was demonstrated for 3D muscular reconstruction using the DPSO method and Dixon images that allowed generation of patient-specific musculoskeletal models.