Validation of high-resolution water-excitation magnetic resonance imaging for quantitative assessment of thin cartilage layers

Objective: To employ a magnetic resonance (MR) imaging technique for quanti tative assessment of thin cartilage layers, and to validate the cartilage v olume and thickness measurements. Methods.. We investigated 10 normal elbow joints (age 20 to 69 years) with a 3D gradient echo sequence with selective water excitation (TR 18 ms; TE 9 ms; FA 25(6), resolution 1x0.25x0.25 mm(2) imaging time 19 min). After int erpolating the image data to a 0.125x0.125 mm2 in-plane resolution, the car tilage plates were segmented, reconstructed in 3D, and the cartilage volume and thickness determined with a 3D Euclidean distance transformation algor ithm, independent of the original section plane. The cartilage volume and t hickness values were compared with CT arthrography and A-mode ultrasound. Results: The mean systematic difference between the elbow cartilage volume obtained from MR imaging and CT arthrography was -0.11% (-6.0 mm(3)) and th e mean random difference 5.7% (314 mm3). Except for the fovea capitis radii , the deviations were not statistically significant (range -7.6 to +11.7%). In the humerus, the mean cartilage thickness (average=1.35 mm) was overest imated relative to GT arthrography (+20.7%/+0.23 mm), and slightly underest imated relative to A-mode ultrasound (-6.0%1-0.05 mm). With few exceptions, there were no significant differences between MRI, CT arthrography and ult rasound in the other joint surfaces of the elbow (random deviations between 0.08 and 0.39 mm). Conclusions: The technique presented can be applied for determining the car tilage volume and 3D thickness in joints with thin cartilage layers with a reasonable degree of accuracy. (C) 2000 OsteoArthritis Research Society Int ernational.