Reliability of cervical muscle volume quantification using magnetic resonance imaging

Area of Science:

• Biomedical Engineering
• Radiology
• Anatomy

Background:

• Magnetic Resonance Imaging (MRI) is crucial for quantifying cervical spine musculature in neck pain patients.
• No prior data existed on the scan-rescan reliability of neck muscle volume measurements.

Purpose of the Study:

• To investigate intra- and inter-rater reliability of cervical muscle volume.
• To assess scan-rescan reliability for cervical muscle volume quantification.

Main Methods:

• Five asymptomatic individuals underwent repeated MRI scans within one hour.
• Two raters manually traced specific neck muscles (levator scapulae, multifidus, semispinalis cervicis, semispinalis capitis, splenius capitis, splenius cervicis, sternocleidomastoid).
• Reliability was assessed using intra-class correlation coefficients (ICCs), Lin's concordance coefficient, and Bland-Altman plots.

Main Results:

• Intra-rater reliability for muscle quantification was excellent (ICCs 0.78–0.96).
• Inter-rater reliability was excellent for sternocleidomastoid (ICC 0.92) and splenius capitis (ICC 0.77), and fair to good for others.
• Scan-rescan reliability was excellent across all muscles (ICCs 0.94–0.98).

Conclusions:

• Reliability issues stem from manual tracing, not scanning protocols.
• The proposed MRI methods are suitable for establishing normative cervical muscle volume data.
• These methods can aid in comparing muscle volumes between individuals with and without neck pain.