Reliability of Evaluating Lumbar Paraspinal Muscles Using Magnetic Resonance and Computed Tomography Imaging in Adults with and without Low Back Pain: A Systematic Review and Meta-Analysis.

OBJECTIVE

To evaluate the reliability of magnetic resonance imaging (MRI) and computed tomography (CT) for assessing lumbar paraspinal muscles (LPMs) in adults with and without low back pain (LBP).

DESIGN

Systematic review and meta-analysis following PRISMA of Diagnostic Test Accuracy.

LITERATURE SEARCH: Searches of Medline, Embase, PubMed Central, Science Citation Index Expanded, Emerging Sources Citation Index, Social Sciences Citation Index, and the KCI-Korean Journal Database were conducted in 2022 and updated in 2024, without date restrictions.

STUDY SELECTION CRITERIA: Studies reporting the reliability of MRI and/or CT for assessing the size and/or quality of the erector spinae, multifidi, quadratus lumborum, and/or psoas in adults, with/without LBP, were included. Two reviewers independently selected the studies.

DATA SYNTHESIS: Studies reporting intraclass correlation coefficients (ICCs) were included in random-effects meta-analyses. All LPMs were analyzed in adults with and without LBP, followed by muscle-specific analyses regardless of LBP status.

RESULTS

Seventy-three MRI and thirteen CT studies were included. Most studies assessed the erector spinae and multifidi at lower lumbar levels and underreported reliability methods. Fifty-seven MRI and nine CT studies were included in meta-analyses. MRI demonstrated excellent reliability for evaluating LPM size and quality in adults with/without LBP (ICCs: 0.922-0.983). CT assessments showed good-to-excellent reliability (ICCs: 0.839-0.994), based on ≤4 studies. Muscle-specific meta-analyses confirmed excellent reliability for MRI (ICCs: 0.931-0.977) and CT (ICCs: 0.929-0.970).

CONCLUSION

LPM assessments using MRI and CT demonstrate excellent reliability regardless of LBP status. However, limited CT evidence and underreported methodology highlight the need for additional research and improved reliability.