Evaluating Paraspinal Back Muscles Using Computed Tomography (CT) and Magnetic Resonance Imaging (MRI): Reliability Analysis and Correlation with Intervertebral Disc Pathology

Purpose: To investigate the reliability of CT and MRI for quantitative and qualitative analyses of lumbar paraspinal muscle fatty infiltration (PSFI) and correlation of PSFI with intervertebral disc pathology. Materials and Methods: Lumbar spine CT and MRI of 36 subjects were reviewed retrospectively. Two observers independently outlined lumbar paraspinal muscles at each mid-intervertebral disc level. Paraspinal muscles on CT and MRI were graded according to the Goutallier grading system (GGS). The area, mean value, and standard deviation (SD) of the Hounsfield unit (HU) were obtained. Intervertebral discs were assessed on axial image of T2WI at each level. Correlations between qualitative and quantitative data and intervertebral disc pathology, age, and sex were evaluated. Results: Inter- and intra-observer agreements for results of GGS on MRI were substantial (κ = 0.79) and moderate (κ = 0.59), respectively. Inter- and intra-observer agreements for results of GGS on CT were almost perfect (κ = 0.88) and substantial (κ = 0.66), respectively. Quantitative measurements of HU showed almost perfect inter- and intra-observer reliabilities (κ = 0.82 and κ = 0.99, respectively). There were statistically significant correlations between intervertebral disc pathology and PSFI at L1-2, L2-3, and L4-5 levels on MRI and at L1-2 and L3-4 levels on CT. Age showed significant correlation with results of GGS at all levels on CT and MRI. Conclusion: This study showed that GGS results and HU measurements could be useful for evaluating PSFI because they showed correlations with intervertebral disc pathology results at certain levels.