Aim: To assess differences in the evaluation of degenerative changes of the lumbar spine MRI among experienced radiologists, radiology residents and other specialists. Method: Nine evaluators from three groups: experienced radiologists (ER), radiology residents (RR) and non-radiologists (NR) independently assessed MRI examinations of the lumbar spine of 55 patients.
All segments of the lumbar spine were evaluated for presence and degree of severity of changes in seven parameters: disc height, disc protrusion or herniation, vertebral body signal changes, spondylolisthesis, dorsal osteophytes, arthritic changes on the intervertebral joints and thickness of ligamenta flava. Statistical evaluation of the degree of agreement among multiple observers was made in individual segments of the lumbar spine by Fleiss kappa; the degree of severity of MRI findings was assessed by Friedman test and pairwise multiple comparisons by Nemenyi post hoc test.
Results: Degree of agreement in the overall MRI comparison in individual segments of the lumbar spine ranged from 0.32 to 0.80 kappa values, most frequently from 0.5 to 0.7 - it was moderate or substantial. Testing of the degree of severity of MRI findings showed a statistically significant differences (p < 0.0001) among individual evaluators.
In each group of evaluators (ER, RR, NR) was always one evaluator who differed significantly from the other two. These differences were smallest among experienced radiologists.
Conclusion: We found statistically significant differences in the assessment of MRI findings among individual evaluators and also within groups of evaluators. These differences were smallest in the group of experienced radiologists.