Quantitative paraspinal muscle measurements: inter-software reliability and agreement using OsiriX and ImageJ

Area of Science:

• Spinal imaging analysis
• Musculoskeletal research
• Medical image processing

Background:

• Paraspinal muscle variations (CSA, composition) are linked to low back pain (LBP) risk and recovery.
• Previous studies show varying reliability in muscle measurements due to diverse protocols and software.
• A standardized protocol for tissue segmentation and software agreement is lacking.

Purpose of the Study:

• To establish a detailed protocol for paraspinal muscle measurement.
• To assess the reliability and agreement of muscle composition measurements using OsiriX and ImageJ software.

Main Methods:

• A measurement reliability study involving 30 individuals with LBP conditions.
• Lumbar MRI scans (T2-weighted axial images) analyzed for multifidus and erector spinae muscles at L4-L5 and S1.
• Measurements repeated twice using OsiriX and ImageJ by a blinded assessor at least 5 days apart.

Main Results:

• High intrarater reliability (ICCs .77-.99 for OsiriX, .78-.99 for ImageJ) for muscle CSA and composition.
• Excellent agreement between OsiriX and ImageJ for muscle composition measurements (ICCs .81-.99).
• Standard error of measurement (SEM) was comparable between the two software programs.

Conclusions:

• The proposed method demonstrates high reliability for assessing paraspinal muscle CSA, composition, and asymmetry.
• Excellent agreement exists between OsiriX and ImageJ for these measurements.
• Findings may not apply to other software; further validation is needed.