Differences in lumbar paraspinal muscle morphology in patients with sagittal malalignment undergoing posterior lumbar fusion surgery

Purpose To investigate whether (1) there is a difference between patients with normal or sagittal spinal and spinopelvic malalignment in terms of their paraspinal muscle composition and (2) if sagittal malalignment can be predicted using muscle parameters. Methods A retrospective review of patients undergoing posterior lumbar fusion surgery was conducted. A MRI-based muscle measurement technique was used to assess the cross-sectional area, the functional cross-sectional area, the intramuscular fat and fat infiltration (FI) for the psoas and the posterior paraspinal muscles (PPM). Intervertebral disc degeneration was graded for levels L1 to S1. Sagittal vertical axis (SVA; >= 50 mm defined as spinal malalignment), pelvic incidence (PI) and lumbar lordosis (LL) were measured, and PI-LL mismatch (PI-LL; >= 10 degrees defined as spinopelvic malalignment) was calculated. A receiver operating characteristic (ROC) analysis was conducted to determine the specificity and sensitivity of the FIPPM for predicting sagittal malalignment. Results One hundred and fifty patients were analysed. The PI-LL and SVA malalignment groups were found to have a significantly higher FIPPM (PI-LL:47.0 vs. 42.1%; p = 0.019; SVA: 47.7 vs. 41.8%; p = 0.040). ROC analysis predicted sagittal spinal malalignment using FIPPM (cut-off value 42.69%) with a sensitivity of 73.4% and a specificity of 54.1% with an area under the curve of 0.662. Conclusion Significant differences in the muscle composition between normal and malalignment groups with respect to FIPPM in both sagittal spinal and spinopelvic alignment were found. This work underlines the imminent impact of the paraspinal musculature on the sagittal alignment.