FLUID-SENSITIVE SEQUENCES MODIFY THE MODIC CLASSIFICATION IN MRI

ABSTRACT Objective: To evaluate whether the magnetic resonance imaging (MRI) acquisition with the Dixon fat suppression technique alters the detection of signal abnormalities in the subchondral bone of the vertebral endplates in the lumbar spine. Methods: retrospective study include patients who underwent lumbar spine MRI with acquisition of T1- and T2-weighted fast spin echo sequences (routine protocol) and the multipoint Dixon sequence with T2 weighting and fat suppression. Two radiologists qualitatively classified Modic-type changes (absent, Modic I, Modic II, and Modic III) independently and blindly. When the subchondral signal abnormality was identified, each radiologist measured the maximum depth in relation to the vertebral plateau. The radiologists’ assessments were used to analyze interobserver agreement and one of the radiologists carried out a second analysis of the exams, enabling intra-observer assessment. Results: The results showed a significant increase in the detection of the edema pattern (Modic I) in qualitative and quantitative analyzes when using the Dixon technique, in comparison to the routine sequences, with an increase in the prevalence and extent of involvement by the bone edema pattern. There was no significant difference in the detection of the adipose replacement pattern (Modic II) between the two techniques. There was moderate to substantial intra-observer correlation for the traditional protocol and substantial to high agreement for the analysis of the Dixon technique images. Conclusions: The Dixon technique showed superiority in the detection of Modic I in relation to the traditional protocol, and allowed the detection, on average, of greater depth of involvement of subchondral bone changes. There was good intra-observer and inter-observer reproducibility in identifying Modic abnormalities. Evidence level III, comparative retrospective study.