MRI versus radiography as eligibility screening tool for knee osteoarthritis clinical trials: Data from the osteoarthritis initiative

Objective

To estimate the prevalence of knee structural pathologies that may warrant exclusion from disease-modifying osteoarthritis drug (DMOAD) trials, based on MRI versus radiography, among participants who would otherwise be considered eligible for enrollment based on commonly used radiographic inclusionary criteria.

Methods

We selected participants from the baseline visit of the Osteoarthritis Initiative that met radiographic structural and clinical eligibility criteria for a DMOAD clinical trial: Radiographic OA (Kellgren-Lawrence grade 2 or 3 with medial minimum joint space width ≥ 1.5 mm); and WOMAC knee pain score between 8 and 18 (0–20 scale). A musculoskeletal radiologist read radiographs and a shortened 3-T MRI protocol for structural pathologies that have been suggested as exclusionary from DMOAD clinical trial participation, as proposed by the Rapid Osteoarthritis MRI Eligibility Score (ROAMES), including meniscal root tears, subchondral insufficiency fractures, osteonecrosis, tumors and bone marrow infiltration, and acute traumatic changes.

Results

Among 380 eligible participants (219 KL2 knees, 161 KL3 knees), structural pathologies that warrant exclusion from clinical trials were found in 38 knees (10 %) based on MRI, versus three knees (0.8 %) on radiography. Based on MRI, the estimated probability of an exclusionary finding in a KL2 knee was 0.06 (95 %CI: 0.03, 0.09), and 0.16 (95 %CI: 0.10, 0.21) in a KL3 knee. The most common finding was complete meniscus posterior root tear (31 knees, 8.2 %).

Conclusion

Structural pathologies that are unlikely to respond to investigational DMOADs were found in ∼10 % of eligible participants based on MRI screening, largely driven by complete meniscal root tears, while these findings were not visible on radiographs.