Objective: The purpose of our study was to evaluate several commonly used MR sequences to determine how accurately each demonstrates the thickness of the articular cartilage of the humeral head.
Materials and methods: Ten cadaveric shoulders (age at death, 58-92 years; mean, 79 years) were imaged with fat-suppressed transaxial T1-weighted spin-echo three-dimensional gradient-recalled sequences, both before and after injection of 12 ml of diluted gadopentetate dimeglumine. Articular cartilage was measured to the nearest 10th of a millimeter on the MR images and corresponding anatomic sections.
Results: Cartilage could not be differentiated from surrounding structures in 14 of 112 locations (13%) on the spin-echo images obtained without contrast material, in 4 of 112 locations (4%) on the spin-echo images obtained with contrast material, in 4 of 112 locations (4%) on the gradient-echo images obtained without contrast material, and in 6 of 112 locations (5%) on the gradient-echo images obtained with contrast material. Mean true cartilage thickness was 1.23 mm (SD, 0.52 mm). The mean MR-anatomic differences (absolute values) were 0.38 mm for the spin-echo images obtained without contrast material, 0.42 mm for the spin-echo images obtained with contrast material, 0.49 mm for the gradient-echo images obtained without contrast material, and 0.37 mm for the gradient-echo images obtained with contrast material. There was a tendency to overestimate thin cartilage and to underestimate thick cartilage.
Conclusion: Several of the routinely used MR sequences, with and without intraarticular contrast medium, may cause errors in the assessment of the articular cartilage of the humeral head. Some of these errors result from either insufficient contrast between cartilage and surrounding structures or inadequately concentrated contrast medium.