Accuracy of cartilage and subchondral bone spatial thickness distribution from MRI

Purpose: To assess three-dimensional measurement accuracy of articular cartilage (AC) and subchondral bone (SB) thickness from MRI. Materials and Methods: A computer program was used to calculate AC and SB thickness from MRI (three-dimensional spoiled gradient echo (SPGR), .31-mm resolution, 1-mm slice thickness) of six adult femoral heads. Specimens were imaged in five anatomical planes ranging between +30degrees to -30degrees from neutral and cut into 2-mm thick sections along the five anatomical planes. Faxitron x-ray was used to produce microradiographic (.05-mm resolution) images of the sections. Results: In-plane measurement accuracy was. 165+/-.108 mm for AC thickness and .387+/-.174 mm for SB thickness. Taking into account chemical-shift misregistration in SB thickness, accuracy of measurements improved to .213+/-128 mm. Out-of-plane (three-dimensional) thickness accuracy of the model, assessed by numerical simulation, was .015 mm. However, three-dimensional thickness errors in specimens were .319+/-.256 mm for AC and .253+/-.183 mm for SB thickness. Conclusion: Errors in three-dimensional AC thickness were attributed to volume-averaging effects caused by oblique intersection of the image plane with the joint surface. Errors in three-dimensional SB thickness were attributed to chemical-shift artifact. We conclude that accuracy of AC thickness is within clinically acceptable standards but that more sophisticated pulse sequences are needed to improve the measurement of SB thickness.