Reliability of SCIseg Automated Measurement of Midsagittal Tissue Bridges in Spinal Cord Injuries Using an External...

Area of Science:

• Neuroimaging
• Spinal Cord Injury Research
• Biomedical Engineering

Background:

• Spinal cord injury (SCI) research requires accurate assessment of lesion damage.
• T2-weighted magnetic resonance imaging (MRI) is crucial for visualizing SCI.
• Manual measurement of midsagittal tissue bridges can be time-consuming.

Purpose of the Study:

• To evaluate the interrater reliability between automated and manual measurements of lesion damage in spinal cord injury (SCI).
• To compare automated (SCIsegV2/SCT) and manual T2-weighted MRI assessments of midsagittal tissue bridges.
• To determine if automated methods can reliably replicate manual assessments for SCI research.

Main Methods:

• Twenty-one T2-weighted MRIs from SCI patients were analyzed.
• Manual measurements of midsagittal tissue bridges were performed by an experienced rater.
• Automated measurements utilized the SCIsegV2 function within the Spinal Cord Toolbox (SCT).
• Interrater reliability was assessed using intraclass correlation coefficients (ICC) and Cohen's kappa.

Main Results:

• Excellent interrater reliability was found for both ventral (ICC 0.94) and dorsal (ICC 0.99) tissue bridges.
• Percentage agreement between manual and automated measures was high (90.8%).
• Substantial agreement was observed using Cohen's kappa for detecting tissue bridges (kappa > 0.79).

Conclusions:

• Automated and manual measurements of midsagittal tissue bridges in SCI demonstrate high reliability.
• Automated MRI analysis offers a reliable and potentially faster method for SCI research.
• This automated approach may expedite studies correlating tissue bridges with functional outcomes in SCI.