Magnetic resonance imaging evaluation of weight-bearing subchondral trabecular bone in the knee

Abstract

Objective: Changes in weight-bearing subchondral bone are central to osteoarthritis (OA) pathophysiology. Using MR, knee trabecular bone is typically assessed in the axial plane, however partial volume artifacts limit the utility of MR methods for femorotibial compartment subchondral bone analysis. Oblique-coronal acquisitions may enable direct visualization and quantification of the expected increases in femorotibial subchondral trabecular bone.

Methods: MR acquisition parameters were first optimized at 3 Tesla. Thereafter, five volunteers underwent axial and coronal exams of their right knee. Each image series was evaluated visually and quantitatively. An anatomically standardized region-of-interest was placed on both the medial and lateral tibial plateaus of all coronal slices containing subchondral bone. Mean and maximum marrow signal was measured, and "bone signal" was calculated.

Results: The MR acquisition had spatial resolution 0.2 × 0.2 × 1.0 mm and acquisition time 10.5 min. The two asymptomatic knees exhibited prominent horizontal trabeculae in the tibial subchondral bone, while the one confirmed OA knee had disorganized subchondral bone and absent horizontal trabeculae. The subchondral bone signal was 8-14% higher in both compartments of the OA knee than the asymptomatic knees.

Conclusion: The weight-bearing femorotibial subchondral trabecular bone can be directly visualized and changes quantified in the coronal-oblique plane. Qualitative and quantitative assessments can be performed using the resultant images and may provide a method to discriminate between the healthy and OA knees. These methods should enable a quantitative evaluation of the role of weight-bearing subchondral bone in the natural history of knee OA to be undertaken.

Fig. 1

The region-of-interest (ROI) conforms to the cartilage-bone interface on the medial tibial plateau. Arrow identifies chemical shift artifact on femoral condyle that does not interfere with subchondral bone analysis

Fig. 2

Bone marrow signal versus flip angle. A flip angle of 45° to 50° maximizes bone marrow signal using a 3D GRE acquisition with a TR of 20 ms at 3 T

Fig. 3

SNR versus flip angle. A flip angle of approximately 50° maximizes bone marrow SNR with a 3D GRE acquisition with a TR of 20 ms at 3 T

Fig. 4

CNR versus flip angle. CNR between bone marrow and muscle did not vary substantially as a function of flip angle (green squares). CNR between bone marrow and subcutaneous fat was minimized at a flip angle of 35° (red triangles)

Fig. 5

a. Sample image from an asymptomatic volunteer (medial side is on right). In these knees, the subchondral bone appeared organized and the contour of the tibial plateau was smooth and concave. b. Sample image from a volunteer with known radiographic knee OA (KL 3 [19]) with medial joint space narrowing. This knee demonstrates disorganized trabeculae and an irregular and flattened tibial plateau contour with thickened subchondral bone

Fig. 6

Medial tibia bone signal topographs

Fig. 7

Lateral tibia bone signal topographs