Abstract
Objectives
To evaluate fat fraction of subchondral bone marrow oedema-like lesions (BMLs) and nearby marrow in knee osteoarthritis (OA) using Dixon MR imaging and to compare conventional volume measurements of BML change with an alternative technique using Dixon MRI and image registration.
Methods
We analyzed 3T MR images from 17 symptomatic knee OA subjects (mean age: 61.3 years). BML volume was measured by manually segmenting sagittal fat-suppressed proton density (PD) weighted images at two visits. In addition, regions of interest (ROIs) from BMLs and nearby non-BML marrow were manually segmented on Dixon water images at either the initial or follow-up visit; image registration of the Dixon images at different time points was performed to allow calculation of median fat fraction from corresponding ROIs at both visits. The fat fraction of BMLs was compared with non-BML marrow using the Wilcoxon signed rank test and Spearman's correlation coefficient was used to assess the association between change in BML volume and change in fat fraction.
Results
Fat fraction within BMLs was significantly lower than that of non-BML marrow (72.5% vs 92.1%, p= 0.003). The change in volume of BMLs over time was negatively correlated with change in fat fraction (ρ= -0.58; p= 0.016).
Conclusions
Estimation of BML fat fraction in patients with knee OA using Dixon imaging combined with image registration provides a potential alternative approach to volumetric BML assessment.
To evaluate fat fraction of subchondral bone marrow oedema-like lesions (BMLs) and nearby marrow in knee osteoarthritis (OA) using Dixon MR imaging and to compare conventional volume measurements of BML change with an alternative technique using Dixon MRI and image registration.
Methods
We analyzed 3T MR images from 17 symptomatic knee OA subjects (mean age: 61.3 years). BML volume was measured by manually segmenting sagittal fat-suppressed proton density (PD) weighted images at two visits. In addition, regions of interest (ROIs) from BMLs and nearby non-BML marrow were manually segmented on Dixon water images at either the initial or follow-up visit; image registration of the Dixon images at different time points was performed to allow calculation of median fat fraction from corresponding ROIs at both visits. The fat fraction of BMLs was compared with non-BML marrow using the Wilcoxon signed rank test and Spearman's correlation coefficient was used to assess the association between change in BML volume and change in fat fraction.
Results
Fat fraction within BMLs was significantly lower than that of non-BML marrow (72.5% vs 92.1%, p= 0.003). The change in volume of BMLs over time was negatively correlated with change in fat fraction (ρ= -0.58; p= 0.016).
Conclusions
Estimation of BML fat fraction in patients with knee OA using Dixon imaging combined with image registration provides a potential alternative approach to volumetric BML assessment.
| Original language | Undefined |
|---|---|
| Article number | 100067 |
| Journal | Osteoarthritis Imaging |
| Early online date | 17 Jun 2022 |
| DOIs | |
| Publication status | Published - 6 Jul 2022 |