Quantitation of articular cartilage using magnetic resonance imaging and three-dimensional reconstruction

• Published in: Journal of orthopaedic research Vol. 13; no. 6; p. 814
• Main Authors: Marshall, K W, Mikulis, D J, Guthrie, B M
• Format: Journal Article
• Language: English
• Published: United States 01.11.1995
• Subjects: Adult; Animals; Cartilage, Articular - pathology; Cattle; Humans; Image Interpretation, Computer-Assisted; Knee Joint - pathology; Magnetic Resonance Imaging - methods; Male; Observer Variation; Phantoms, Imaging; Reproducibility of Results
• ISSN: 0736-0266; 1554-527X
• DOI: 10.1002/jor.1100130603

A quadrature knee coil was used in conjunction with a magnetic resonance imaging scanner for quantitation of test phantom volumes, ex vivo bovine cartilage thickness, and in vivo human articular cartilage volumes. Optimal magnetic resonance parameters were obtained by testing a series of spin-echo and gradient-echo pulse sequences to determine the sequence that provided the highest resolution of articular cartilage and best defined the cartilage interfaces with synovial fluid and subchondral bone. Extensive testing revealed that two sequences were required to define articular cartilage accurately: a spoiled gradient-echo sequence and a steady state free-precession sequence. Three-dimensional reconstruction and statistical analyses of test phantoms and of bovine and human cartilage images were performed. Differences between actual phantom volumes and three-dimensional measurements demonstrated that, as magnetic resonance slice thickness was increased, the measurement variability also increased (coefficient of variation ranging from 1.7 +/- 1.3% for 1.0 mm slice thickness to 22.7 +/- 1.9% for 3.0 mm slice thickness). When the phantom volume was greater than 1,800 mm3, the intraobserver, interobserver and interscan accuracies were greater than 97, 98, and 96%, respectively. This high degree of reproducibility pertained for the data on in vivo human cartilage data also. For experienced observers, the intraobserver and interobserver reproducibility were greater than 98 and 97%, respectively. The interscan reproducibility was greater than 98%. These data demonstrate that improved magnetic resonance pulse sequencing, in conjunction with three-dimensional reconstruction and measurement techniques, can accurately and reproducibly measure the volume of articular cartilage. Clinical application of this approach offers the potential for early diagnosis of osteoarthritis and for serial, noninvasive assessment of changes in articular cartilage volume in response to therapeutic modalities.