Abstract
Aim Post-traumatic osteoarthritis and osteochondral injuries can cause significant pain and morbidity. Appropriate MRI sequences combined with image analysis techniques can be used to reproducibly measure quantitative cartilage parameters, hence offering a tool for monitoring and detecting degenerative change earlier than previously possible. We demonstrate the performance of a directional gradient vector flow (dGVF) snake segmentation algorithm on an isotropic MR sequence, which allows segmentation of the full articular surfaces (including malleoli) of the ankle.
Method Eight ankles were imaged using a 1.5T MRI scanner with an isotropic 3D T1 weighted FLASH sequence with water excitation, resolution 0.3 x 0.3 x 0.3 mm. A subset of five ankles were imaged four times with repositioning and re-shimming of the magnet between acquisitions. Images were interpolated to 0.15 mm3 and segmented using a dGVF snake. Following 3D reconstruction of the cartilage layers normal thickness from cartilage to bone was measured at each voxel on the cartilage surface.
Results The mean cartilage thickness (±S.D) was 1.80 mm (±0.05 mm); 1.83 mm (±0.07 mm) and 1.81 mm (±0.07 mm) for the talus, tibia and cumulative ankle cartilage respectively. To measure the technical precision of the segmentation method we determined the coefficient of variation of the four repeated measurements in five ankles. The mean coefficients of variation (min-max) from the repeated measurements were 1.74% (0.69%–3.57%); 1.20% (0.26%–3.06%) and 1.52% (0.26%–3.57%) for the talus, tibia and cumulative ankle cartilage respectively.
Conclusion We believe that the reported isotropic image sequence and segmentation algorithm is a valid tool for quantitative assessment of the entire ankle joint. A possible application is the early detection of cartilage injury and degenerative change due to injury or illness.
Correspondence should be addressed to BOFSS, c/o Wrightington Wigan and Leigh NHS Trust, Hall Lane, Appley Bridge, Wigan, Lancashire WN7 9EP.
