Objectives: Although accurate spatial measurement of cartilage thickness from MRI is possible, no studies have assessed the accuracy of measuring cartilage defect dimensions from MRI. In addition, current MR grading scales for assessing cartilage lesions have limited categories, and little is known about how well these scales correlate with histological assessment of the lesion. The objective of this preliminary study is to address both these issues.
Methods: We performed two experiments on four cadaver knee joints from elderly donors: Experiment 1 assessed the accuracy of measuring controlled defects in cartilage, and Experiment 2 compared MRI grading (Noyes scale) of natural cartilage lesions to histological grading (Mankin scale) of the sectioned cartilage tissue. MRI was performed on 1.5 T clinical scanner (fat-suppressed 3D-SPGR at TR/TE/alpha=55/13.5/45 and 256 x 256 matrix).
Results: The mean difference between defect diameters measured and introduced was less than 0.1mm, which was statistically insignificant (P=0.754). Defect depth was less accurate at >0.4mm, significantly under predicting actual defect depth (P=0.004). Correlation between Noyes grading scores and Mankin grading scores of natural lesions was moderately high (r=0.7) and statistically significant (P=0.001).
Conclusions: Three-dimensional mapping of cartilage thickness shows great promise for the accurate measurement of focal cartilage defects, though improvement is needed. The Noyes grading scale is consistent with histological Mankin grading of cartilage lesions, though enhancement of MR grading scales is needed, and warranted, based on the signal intensity information available from clinical MRI. Integration of these two analyses-focal defect measurement and signal intensity analysis-could potentially result in a valuable clinical tool for early osteoarthritis diagnosis and longitudinal tracking.