Water distribution patterns inside bovine articular cartilage as visualized by1H magnetic resonance imaging

Abstract

Objective To develop a magnetic resonance imaging (MRI) technique to non-invasively map water volume fraction (WVF) in articular cartilage. Special emphasis was placed on spatial resolution and temporal considerations, aimed at creating a procedure feasible for eventual human studies.DesignAbsolute proton density MR images of intact, ex vivo bovine patellae were calculated from fullyT1 relaxed, short echo time images. This was accomplished through compensation for T2decay with calculated T2maps. Calibration of the signal intensity in the image was accomplished with the use of H2O:D2O phantoms, where the WVF was varied from 0.95 to 0.75. Application of the calibration curve to the entire image yielded images that represent WVF on a pixel by pixel basis. Calculations of water content by weight were performed by considering the density of the solid content.Results Using four echo time points, experiments comparing MR images from single-echo and multi-echo spin echo sequences yielded similar results. T2decreased with depth through the cartilage, with a maximum at the articular surface of approx 100ms, and a ∼50ms minimum at the bone/cartilage interface. The WVF through the depth of the cartilage showed a similar trend, decreasing from 0.9 at the surface, to 0.7 at the bone/cartilage interface. Translation to a weight percent yielded ∼86% weight at the surface, trending down to ∼63% at the bone/cartilage interface, with an average of 74.5% for five patellae. These MRI derived values were compared to the measured weight of water in excised cartilage plugs from the same patellae and showed remarkably close agreement.Conclusion We have demonstrated that MRI can non-invasively map WVF in cartilage in a pixel by pixel manner. This was accomplished in a time span that was clinically feasible, allowing the routine use of this method in a clinical setting. Moreover, this procedure employed standard MRI equipment and pulse sequences, avoiding the need for hardware modifications and using simple post processing methods. However, baseline studies need to be performed prior to incorporation into a standard radiological evaluation. Implications in the diagnosis of osteoarthritis (OA) are discussed.