Development and reliability of a systematic method to evaluate lumbar paraspinal muscle size and quality in computed tomography images.

BACKGROUND

Lumbar paraspinal muscles (LPMs) are critical for spinal stability, and their morphological alterations are associated with chronic low back pain (cLBP). Computed tomography (CT) offers a feasible alternative to magnetic resonance imaging for assessing LPM morphology; however, its broader use has been hindered by inconsistent measurement methods and limited reliability data. This study aimed to describe a systematic CT-based method for evaluating LPM cross-sectional area (CSA) and attenuation (MA) in adults with cLBP and to determine its intra and inter-examiner procedural reliability.

METHODS

High-resolution CT images were used to analyze bilateral LPMs: erector spinae (L1-L5), multifidi (L1-S1), psoas (L3), and quadratus lumborum (L3). Two trained examiners independently performed image selection and analysis, blinded to subject identity and prior results. Procedural reliability was quantified using intraclass correlation coefficients (ICC2,1) with 95% confidence intervals (CI). Mean ICCs<0.50 indicated poor, 0.50-0.74 moderate, 0.75-0.90 good, and >0.90 excellent procedural reliability.

RESULTS

Thirty-six participants (40% female; age 48±15 years; BMI 27.2±4.3 kg/m²) with moderate disability were included. Intraexaminer procedural reliability was excellent for CSA of the erector spinae, psoas, and quadratus lumborum (ICC point estimates >0.916, >0.959, and >0.962, respectively); and good to excellent for multifidi CSA (ICC point estimates: 0.770-0.916), with greater variability at L1 and L2. Inter-examiner procedural reliability was excellent for CSA of the erector spinae, psoas, and quadratum lumborum (ICC point estimates >0.916, >0.946, and >0.941, respectively); and moderate to excellent for multifidi CSA (ICC point estimates: 0.593-0.945), with lower reliability at upper lumbar levels. Intra and inter-examiner procedural reliability for MA was excellent across all LPMs.

CONCLUSIONS

The presented CT-based protocol provides a standardized and reproducible method for assessing LPM CSA and MA in adults with cLBP. Demonstrated high procedural reliability-including image selection and analysis-supports its use for consistent evaluation of LPM morphology in both clinical and research contexts, particularly where muscle degeneration complicates image-based assessment.