Objective: Our objective was to analyze MRI contrast-enhancement patterns in arthritic and nonarthritic knees and the relationship of those patterns with clinical, laboratory, and histologic synovium markers.

Materials and methods: Dynamic contrast-enhanced MRI was performed in nine arthritic and three nonarthritic knees of juvenile rabbits. A two-compartment pharmacokinetic model of signal intensity-time data was implemented to generate parametric maps of signal slope, maximal percentage of signal change, capillary permeability, leakage space volume, and time-to-peak. MRI values were compared with clinical, laboratory, and histologic markers for evaluation of synovial changes during the progression of arthritis.

Results: Parametric maps of capillary permeability and signal slope depicted significant differences between arthritic and nonarthritic knees. Arthritic knees showed increased capillary permeability (p = 0.006) and signal slope (p = 0.01) with time after onset of disease as opposed to nonarthritic knees (permeability, p = 0.65; slope, p = 0.56). Significant correlations were found between temporal changes in capillary permeability (p = 0.002), signal slope (p = 0.003), and serum concentrations of amyloid A. No relationship was noted between any MRI parameters and histologic scores. The discriminative power of MRI indexes varied according to the stage of arthritis: time-to-peak was most accurate for differentiation of presence versus absence of arthritis in early arthritis (day 1, p = 0.0002), and signal slope was most accurate in midterm arthritis (day 14, p = 0.001).

Conclusion: In vivo capillary permeability and signal slope have distinctive dynamic MRI properties. The accuracy of MRI parameters for diagnostic evaluation of experimental arthritis differs according to the stage of disease.