AbstractBackground Vertebral endplate bone marrow lesions ("Modic changes", MC) are a putative cause of chronic low back pain in some individuals. However, associations between MC subtypes and clinical symptoms are variable, which limits the utility of conventional imaging in diagnosing and treating chronic low back pain patients with MC. Purpose Develop a quantitative imaging biomarker for measuring compositional abnormalities in MC-affected endplates using water-fat MRI and test for associations with patient-reported outcomes, conventional MC type classification, and longitudinal conversion of MC types. Design/Setting Single center prospective longitudinal observational cohort study. Patient sample 125 adults with non-specific chronic low back pain, 59 of whom (47%) had radiologist-diagnosed lumbar MC at baseline. Follow-up imaging (12-months) was collected for 37 participants (14 with MC). Outcome Measures A novel measure of MC composition from water-fat MRI, defined as the percentage of the endplate region with abnormally hypo-fatty or hyper-fatty bone marrow. Conventional evaluation of MC presence/absence and type, and ordinal scores for MC height and area as evaluated by a musculoskeletal radiologist on T1- and T2-weighted MRI. Patient-reported outcome measures included the Oswestry Disability Index; the Pain, Enjoyment, and General Activity scale; the visual analogue scale; and the PROMIS physical function and pain interference scales. Methods Multivariable linear regression models adjusted for age, sex, and MC count were used to test for associations between patient-reported outcomes and 1) water-fat MRI biomarker values; and 2) conventional MC parameters (presence/absence, type, height, area). Cohen's f2 was used to evaluate effect size (f2 ≥ 0.02: small effect, f2 ≥ 0.15: medium effect, f2 ≥ 0.35: large effect). Linear regression with a Tukey post-hoc test was used to test for differences in bone marrow composition (water-fat MRI) within and between MC types. Longitudinal changes in biomarker values were compared with longitudinal changes in MC type classification. Results There was substantial heterogeneity in marrow composition among MC-affected endplates, including within lesions of the same MC type. Independent of MC type, more extensive hypo-fatty involvement was associated with worse pain and disability across multiple outcome measures (ODI: p = 0.0007, Cohen's f² = 0.26; PEG, p = 0.026, Cohen's f² = 0.11; VAS, p = 0.018, Cohen's f² = 0.12; PROMIS physical function, p = 0.018, Cohen's f² = 0.12), whereas the extent of hyper-fatty involvement was not associated with patient-reported outcomes (p > 0.41). In contrast, conventional MC type classification was not associated with patient-reported outcomes (p > 0.25). Longitudinally, large changes in marrow composition were observed in individual lesions without corresponding changes in MC type classification. Conclusions Water-fat MRI biomarkers capture pathoanatomic variation in MC-affected marrow that associate more strongly with patient-reported pain, disability, and physical function than conventional MC type classification. Greater hypo-fatty involvement, indicating more severe fibrovascular/edematous changes, is the compositional feature most closely linked to worse symptoms. These biomarkers could improve patient stratification for treatments for vertebrogenic pain.
Bonnheim et al. (Fri,) studied this question.