The predictive value of psoas and paraspinal muscle parameters measured on MRI for severe cage subsidence after standalone lateral lumbar interbody fusion

The effect of psoas and paraspinal muscle parameters on cage subsidence after minimally invasive techniques, such as standalone lateral lumbar interbody fusion (SA-LLIF), is unknown. This study aimed to determine whether the functional cross-sectional area (FCSA) of psoas and lumbar spine extensor muscles (multifidus and erector spinae), and psoas FCSA normalized to the vertebral body area (FCSA/VBA) differ among levels with severe cage subsidence after SA-LLIF when compared to levels without severe cage subsidence. Retrospective single center cohort study. Patients who underwent SA-LLIF between 2008 and 2020 for degenerative conditions using exclusively polyetheretherketone (PEEK) cages, had a lumbar magnetic resonance imaging (MRI) scan within 12 months, a lumbar computed tomography (CT) scan within 6 months prior to surgery, and a postoperative clinical and radiographic follow-up at a minimum of 6 months were included. Severe cage subsidence. MRI measurements included psoas and combined multifidus and erector spinae (paraspinal) FCSA and FCSA/VBA at the L3-L5 pedicles. Following manual segmentation of muscles on axial T2-weighted images using ITK-SNAP (version 3.8.0), the FCSA was calculated using a custom written program on Matlab (version R2019a, The MathWorks, Inc.) that used an automated pixel intensity threshold method to differentiate between fat and muscle. Mean volumetric bone mineral density (vBMD) at L1/2 was measured by quantitative CT. The primary endpoint was severe cage subsidence per level according to the classification by Marchi et al. Multivariable logistic regression analysis was performed using generalized linear mixed models. All analyses were stratified by biological sex. 95 patients (45.3% female) with a total of 188 operated levels were included in the analysis. The patient population was 92.6% Caucasian with a median age at surgery of 65 years. Overall subsidence (Grades 0-III) was 49.5% (53/107 levels) in men versus 58.0% (47/81 levels) in women (p=.302), and severe subsidence (Grades II-III) was 22.4% (24/107 levels) in men versus 25.9% (21/81 levels) in women (p=.608). In men, median psoas FCSA and psoas FCSA/VBA at L3 and L4 were significantly greater in the severe subsidence group when compared to the non-severe subsidence group. No such difference was observed in women. Paraspinal muscle parameters did not differ significantly between non-severe and severe subsidence groups for both sexes. In the multivariable logistic reg