Meeting Banner
Abstract #0433

Classification of OARSI Scored Human Articular Cartilage Explants Through Multiparametric MRI Analysis at 3T

Vanessa A. Lukas1, Kenneth W. Fishbein1, Ping-Chang Lin2, Michael Schr3, Corey P. Neu4, Richard G. Spencer1, David A. Reiter5

1Magnetic Resonance Imaging and Spectroscopy Section, National Institute on Aging, National Institutes of Health, Baltimore, MD, United States; 2Department of Radiology, Howard University College of Medicine, Washington, DC, United States; 3Philips Healthcare, Highland Heights, OH, United States; 4Weldon School of Biomedical Engineering, Purdue University, West Lafayette, IN, United States; 5Clinical Research Branch, National Institute on Aging, National Institutes of Health, Baltimore, MD, United States

In this study, we extend our previous work on classification of degradation of bovine nasal cartilage, performed at 9.4T and 4 C, to the problem of classification of osteoarthritic human articular cartilage explants at physiological temperature (37 C) and at the clinical field strength of 3T. We compared the conventional univariate classification scheme to multiparametric classification according to multivariage Gaussian models. Using the MRI parameters T1, T2, T2*, and ADC, multiparametric classification showed substantially improved accuracy in discriminating between normal and OA cartilage, as determined based on OARSI histological scores. These initial results represent a promising step towards diagnosis of cartilage matrix degradation during early OA using acquisition schemes that are readily available on clinical MRI systems.

Abstract PDF Presentation Video

Keywords

according accuracies accuracy achieved acquisition activity advances aging analyses analyze applicable approaches approved articular best biomedical board body bone bovine cartilage classification classifier classifiers clinical collagen combination combinations comparable conditions containing control covering cross custom datasets define degradation degraded dependence depth detection determined development discriminate disruption early engineering entire evidently example excluding expect extends field froze functions grades harvested health heights histological human implemented improve improved improvements independent initial institute institutes institutional intensity interventions joints knee known limitation look maps matched medicine mixture model models monitoring multivariate nasal national negatives noninvasive note osteoarthritis overlap overlapped ping pixels plugs preparation previous previously progression promising quantitative radiology random rapid readily readout remaining reported represent representing respectively review sample samples school score scored scores scoring section selection sense sensitivity severity sigma slices specificity spectroscopy spin stages strength structure studies subject subjects substantial substantially system table temperature thick tissue training true univariate validation variate west