Diffusion Rate and Near-Equilibrium Distribution of MRI and CT Contrast Agents in Articular Cartilage
+1,2Silvast, T S; 2Kokkonen, H T; 4Nieminen, M T; 1Lammi, M J; 3Quinn, T; 2Jurvelin, J S; 1,2Töyräs, J
Kuopio University Hospital, Kuopio, Finland, 2University of Kuopio, Kuopio, Finland, 3McGill University, Montreal, Quebec, Canada, 4University of
Oulu, Finland
Senior author [email protected]
1
INTRODUCTION
In osteoarthritis, the amount and distribution of collagen,
proteoglycans (PGs) and fixed charge density (FCD) in cartilage are
known to change. The delayed gadolinium enhanced magnetic resonance
imaging of cartilage (dGEMRIC) is a clinical technique for imaging
tissue PG content and distribution [1]. The dGEMRIC is based on
application of anionic contrast agent gadopentetate, which is assumed to
distribute inversely to spatial FCD. Provided that a mobile ion is not
bound to the matrix and the matrix is not restraining the molecule
motion, the equilibrium distribution of mobile anions is controlled by the
distribution of fixed charges in the tissue (Donnan theory) [2]. The
recent papers [3, 4] have reported slow diffusion rates for common
clinical contrast agents gadopentetate and ioxaglate. As the times
allowed for diffusion are considerably shorter in dGEMRIC, true
diffusion equilibrium may not exist at the time of imaging. Thus, the
prediction of FCD distribution may be compromised.
In present study, we apply the contrast enhanced cartilage
tomography (CECT) to determine the effect of molecular size and
charge on the diffusion rate and equilibrium distribution of several
contrast agents, controlling the contrast agent penetration either through
the cartilage surface or deep cartilage.
METHODS
Fresh bovine patellae (n = 5) were prepared within 6 hours after
slaughtering. Eight full thickness (average thickness = 1.95 mm)
cartilage discs (d = 4.0 mm) were detached from each patella. Four
different contrast agents were applied. Two adjacent cartilage disks from
each patella were immersed in each contrast agent. The samples were
mounted in a custom made sample holder so that the contrast agent
diffusion was allowed through the articular surface or through the deep
cartilage. The diffusion process was imaged during the immersion for
29h in 50 ml of phosphate buffered saline (PBS) containing 21 mM
anionic (-1, 1269 g/mol)) ioxaglate (HexabrixTM, Mallinckrodt, St.
Louis, MO, USA) or anionic (-2, 548 g/mol) 100 mM gadopentetate
(Magnevist®, Bayer Schering Pharma, Berlin, Germany) or 180 mM
non-ionic (574 g/mol) gadodiamide (OmniscanTM, Amersham Health,
Oslo, Norway) or 134 mM anionic (-1, 127 g/mol) iodine (i.e.
dissociated NaI) (Sigma-Aldrich, St. Louis, MO, USA). Contrast agent
concentrations were chosen to provide a similar signal-to-noise ratio.
Each solution was adjusted to have isotonic osmolarity (280-330 mOsm)
and pH of 7.4.
The contrast agent distribution maps were determined with CECT
before immersion, and after 1h, 5h, 9h, 16h, 25h, and 29h of immersion
using a clinical peripheral quantitative computed tomographic (pQCT)
instrument (Stratec Medizintechnik GmbH, Pforzheim, Germany). The
tube voltage of 58.0 kVp, slice thickness of 2.3 mm, and pixel size of
0.20x0.20 mm2 were applied. To minimize background noise each slice
was imaged five times and averaged. The pQCT images were analysed
using Matlab (MathWorks Inc., Natick, MA, USA).
RESULTS
With large molecules (ioxaglate, gadopentetate and gadodiamide) it
took over 16h for diffusion to reach the near-equilibrium. In contrast,
diffusion of I- was significantly faster reaching the near-equilibrium
within few hours (Fig. 1). At near-equilibrium, the concentrations of Iand gadopentetate in cartilage were 52.7% and 35.9% of that in the
immersion medium (full thickness average). Contrast agent distribution
profiles at near-equilibrium were significantly different (or even
inverse), depending on the diffusion direction. Furthermore, when the
diffusion was allowed through the deep cartilage the full thickness
average normalized concentration of contrast agent (gadopentetate and
ioxaglate) was significantly lower, as compared to the normalized
concentration recorded when diffusion was allowed through the articular
surface. Moreover, the diffusion was significantly slower through the
deep tissue than through the superficial tissue.
Figure 1. Average distribution profiles (n = 5) of contrast agents at
different time points. The diffusion direction was through the cartilage
surface (Left column) or through the deep cartilage (Right column). 0
and 1 denote cartilage surface and deep cartilage, respectively.
Table 1. Normalized concentration of contrast agent (% of the
immersion solution concentration) in cartilage at different time points
during diffusion (iox = ioxaglate, gadop = gadopentetate, gadod =
gadodiamide, iodine = I-). The contrast agent diffusion direction is
indicated (surf = through articular surface, deep = through deep tissue).
1h
5h
9h
16 h
25 h
29 h
gadop
(surf)
18.5
26.4
30.4
32.9
38.5
35.9
gadop
(deep)
10.5
20.7
26.2
30.1
31.2
31.5
iox
(surf)
24.3
35.6
39.7
41.0
41.9
43.9
iox
(deep)
7.5
15.9
19.3
23.3
27.4
28.2
gadod
(surf)
29.2
45.3
48.8
51.4
53.3
54.9
gadod
(deep)
20.2
41.6
49.5
52.5
54.1
55.9
I(surf)
42.1
49.0
51.9
51.7
51.4
52.7
I(deep)
31.1
43.0
46.8
49.6
50.9
52.2
DISCUSSION
The diffusion rate was significantly dependent on the molecular size.
The >16h equilibration times of large molecules suggest that the Donnan
equilibrium may not be reached when short times, typically in use in
clinical diagnostics, are applied. In some cases the equilibrium might not
have been achieved even after 29h. Then, the direct quantification of
FCD distribution may not be accurate. The normalized concentrations of
contrast agents after 29h were significantly different between the applied
anionic agents. Differences in contrast agent distribution profiles at nearequilibrium suggests that the equilibration does not follow only the
Donnan theory, but other factors, e.g. steric hindrance of diffusion
particularly at deep tissue with higher solid content, may control the
distribution. On the other hand, binding of the solutes to the matrix can
artificially increase the normalized concentration. Thus, the accuracy of
the determination of FCD using the dGEMRIC or CECT techniques may
be questioned. The present contrast agent concentrations were
significantly higher compared to those applied clinically, which may
possibly affect diffusion rates and the near-equilibrium distribution.
Interestingly, the diffusion was found to be significantly slower through
the deep tissue than through the superficial tissue This is in line with the
earlier investigations and suggests that the diffusion rate depends on the
tissue composition and properties [5].
REFERENCES [1] Bashir et al. 1996; [2] Maroudas 1979; [3] Silvast
et al. 2008; [4] Kallioniemi et al. 2007; [5] Evans and Quinn 2005.
Paper No. 241 • 55th Annual Meeting of the Orthopaedic Research Society