Magnetic resonance elastography (MRE) has been used to successfully characterize the mechanical behavior of healthy and diseased muscles, but no study has been performed to investigate the reliability of MRE on lumbar muscles. The objective of this work was to determine the reliability of MRE techniques on lumbar muscles in both ex vivo phantom and in vivo human studies. In this study, fresh porcine leg muscles were used in the phantom study, and 80 healthy adults (38.6 ± 11.2 years, 40 women) were recruited in the human study. Five repeated stiffness maps were obtained from both the phantom and human muscles by using a gradient-echo MRE sequence with a pneumatic vibration on a 1.5 T MR scanner. The technical failure rate, coefficient of variation (CV), and quality score were assessed to evaluate the reliability of MRE, respectively. Analysis of variance was performed to compare the stiffness between different lumbar muscles, and the difference was significant if p < 0.05 after Bonferroni correction. The results showed that the MRE achieved a zero technical failure rate and a low CV of stiffness (6.24 ± 1.41%) in the phantom muscles. However, in the human study, the MRE exhibited high CVs of stiffness (21.57%−25.24%) in the lumbar muscles, and the technical failure rate was higher in psoas muscles (60.0−66.3% in) than in paraspinal muscles (0.0−2.5%). Further, higher quality scores were noticed in paraspinal muscles (7.31−7.71) than those in psoas muscles (1.83−2.06). In conclusion, the MRE was a reliable technique to investigate the mechanical property of lumbar muscles, but it was less reliable to assess stiffness in psoas muscles than paraspinal muscles.
Keywords: MR elastography; lumbar spine; muscle; quality score; stiffness; technical failure rate.