Timothy John Mosher1, Zheng Zhang2,
Ravinder Reddy3, Sanna Boudhar4, Barton Milestone5,
William Morrison6, Kenneth Kwoh7, Felix Eckstein8,
Walter Witschey3, Ari Borthakur3, Cynthia Olson4
1Radiology, Penn State University,
Hershey, PA, United States; 2Brown University; 3University
of Pennsylvania; 4ACRIN; 5Fox Chase Cancer Center; 6Thomas
Jefferson University; 7University of Pittsburgh; 8Paracelsus
Private Medical University
Reproducibility
of morphometric, T1rho, and T2 cartilage biomarkers was evaluated in a
multi-site, multi vendor clinical trial.
53 subjects were evaluated during 4 separate MRI sessions to calculate
intraclass correlation coefficient (ICC). MRI measurements of cartilage
morphometry are highly reproducible in a multi-center/multi-vendor
trial. Subregional T1rho analysis has
poor reproducibility. Improved
reliability is obtained when T1rho analysis is performed using data obtained
using the full thickness of the cartilage, allowing analysis at the level of
the cartilage plate. Cartilage T2 mapping is sufficiently reproducible to
allow for subregional analysis based on depth from articular surface.