Inter-rater reliability between radiologists and orthopaedic surgeons when assessing lumbar degenerative disc disease...

The Pfirrmann grading system is widely used for assessing degenerative disc disease (DDD) on T2-weighted MRI but is criticised for its subjectivity. This study evaluates inter-rater reliability between radiologists and orthopaedic surgeons using this system.

A retrospective review of 60 lumbar spine MRI scans. Four radiologists and four orthopaedic surgeons (two consultants, two specialist registrars per group) independently graded five intervertebral discs per scan using both the 5-level and 8-level Pfirrmann grading systems. Grading was repeated after two months to assess intra-rater reliability.

The highest inter-rater agreement was observed among orthopaedic consultants, with quadratic weighted kappa values of 0.89 (95 % CI: 0.86-0.92) and 0.97 (95 % CI: 0.96-0.98) for the 5- and 8-level grading systems, respectively. Overall agreement across all assessors using the 5-level system was good according to Krippendorff's alpha (ordinal) of 0.67 (95 % CI: 0.63-0.72) and Gwet's AC2 (quadratic) of 0.67 (95 % CI: 0.64-0.71).In contrast, Fleiss's kappa indicated only fair agreement overall, with a coefficient of 0.37 (95 % CI: 0.27-0.47). Radiology consultants achieved the highest intra-rater agreement (76 %) on reassessment when using the 5-level grading system, although they also demonstrated notable grading discrepancies.

This study found that variability exists across assessor groups when comparing inter-rater reliability between radiologists and orthopaedic surgeons using the Pfirrmann grading system for lumbar DDD. This highlights the need for standardised grading and the potential role of quantitative MRI biomarkers to improve reliability.