Objective: In this longitudinal MR study the early stages of joint pathology in two surgically-induced rabbit models of osteoarthritis (OA) were monitored by in vivo contrast-enhanced MRI at 7.1T. Qualitative and quantitative MR data were compared with macroscopic and microscopic findings.
Method: Scanning of mature, male New Zealand White rabbits (N=12) was performed before surgery, and at 2, 4, and 8 weeks after unilateral transection of the anterior cruciate ligament (ACLT), medial meniscectomy (ME), or sham operation. MR-images were simultaneously obtained of both knee joints after intravenous injection of Magnevist. We implemented a 2D T1-weighted (T1w) coronal, fat-saturated gradientecho protocol (68 x 138 microm2, slice thickness 1 mm). Additionally, consecutive 3D gradientecho images were obtained from two sham-operated and two rabbits of the ME group (234 x 273 x 234 microm(3)). ACLT animals were sacrificed at 2 weeks (N=1), and 8 weeks (N=3), ME animals were sacrificed at 4 weeks (N=2), and 8 weeks (N=4), and sham-operated animals were sacrificed at 2 weeks (N=1) and 8 weeks (N=1), respectively.
Results: Both OA models reflected important characteristics of the clinical picture of OA. With MR we were able to monitor time dependently the decline of synovial effusion and the formation of osteophytes. Morphologic MR examination showed a moderate to high accuracy for detecting synovial effusion (75%), meniscus (86%) and cruciate ligament (91%) lesions, and osteophytes (88%) as assessed by macroscopic examination. False-negative MR findings for gross macroscopic changes were due to the relative high slice thickness in 2D scans and the fact that the slices only covered the main weightbearing area of the femorotibial joint. Contour abnormalities of articular cartilage were not reliably detected. Quantitative analysis revealed a statistically significant increase of cartilage signal intensity in medial tibial cartilage (48+/-9% ACLT, and 29+/-9% ME in 2D datasets) as compared to contralateral control knees in two-week scans. Signal enhancement persisted or increased at later dates.
Conclusion: With high-resolution contrast-enhanced MRI at 7.1T the time course of gross pathologic changes in rabbit knees with surgically induced OA can be monitored. Still insufficient spatial resolution and image contrast of the applied 2D protocols limit the sensitivity and prohibit detection of articular cartilage contour abnormalities. However, signal alterations in the cartilage layer indicate alterations of tissue composition at a very early stage of OA development. When used with 3D protocols, contrast-enhanced MRI offers a promising tool for qualitative and quantitative in vivo monitoring of OA in rabbit models.