Responsiveness of a semi-automated novel method of measuring cartilage loss in knee osteoarthritis over two years using 3.0T DESS 3D MRI

Abstract

s / Osteoarthritis and Cartilage 20 (2012) S54–S296 S222 Further, no differences in the mean or SD signal intensity in any of the muscle regions were noted at baseline (data not shown). Changes in ACSAs over 2 years also did not significantly differ between progressor and non-progressor knees (Table 1). When stratifying comparisons for cases with little baseline pain (WOMAC score of 0-1; n1⁄47) vs. those with a score of 5-8 (n1⁄48), the findings were similar (data not shown). Conclusions: Progressor (case) and non-progressor (control) knees in this study were carefully selected from a large subsample, based on two independent measures of structural OA progression (MRI cartilage loss and JSW reduction in X-rays). Further, cases and controls were carefully matched for measures known to be associated with progression or with muscle area. The results of this exploratory study do not provide support that, once radiographic knee OA is established, baseline or longitudinal changes in thigh muscle ACSAs or strength predict (or are associated) with structural progression. 438 EVALUATION OF THE DEPENDENCY OF GLYCOSAMINOGLYCAN (GAG) CHEMICAL EXCHANGE SATURATION TRANSFER (GAGCEST) IMAGING ON CARTILAGE GAG CONTENT IN THE ANKLE AT 3 T B. Schmitt, M. Brix, J. Hofstaetter, R. Windhager, S. Trattnig, S. Domayer. Med. Univ. of Vienna, Vienna, Austria Purpose: This study was performed to evaluate the feasibility of gagCEST imaging in the ankle on a clinical 3-Tesla MR scanner. The dependency of gagCEST signal on cartilage GAG content was investigated by comparison of MRI data with quantitative biochemical assessment of cartilage GAG content. Methods: The study comprised 7 ankle samples from human cadavers, which were examined on a clinical 3 T MR System with a standard knee coil. PDw were acquired with turbo spin-echo (TSE) imaging and fatsat (FS) in the sagittal plane (TE1⁄426ms, TR1⁄44000ms, resolution1⁄40.4x0.4x3mm3). GagCEST imaging was performed using a segmented 3D RF-spoiled gradient-echo (GRE) sequence (TE1⁄43.49ms, TR 1⁄49.1ms, resolution1⁄40.6x0.6x3.3mm3, scan time 10:30 min). Selective RF presaturation was achieved using a series of 3 Gaussian RFPulses with pulse duration sp1⁄4100ms, an interpulse delay sd1⁄410ms and a B1 of 2.6mT. Z-spectra from images were corrected for B0 inhomogeneities on a pixel-by-pixel basis by a smoothing spline method. The asymmetry of the magnetization transfer rate (MTR) as determined by MTRasym (d) 1⁄4 MTR(+d)-MTR(-d) was integrated over the offset range from 0.5 2ppm, which corresponds to the resonance signal distribution from exchangeable GAG -OH protons, and used as signal intensity for gagCEST images. For quantitative biochemical analysis of absolute GAG content in cartilage, as gold standard, the tibial and talar cartilage compartments were divided into three segments (lateral, central, medial) with 1cm width in the sagittal plane (Fig. 1a). In each segment, 5 contiguous cartilage samples were taken, and a GAG assay (Blyscan B3000 GAG Assay) was used to determine absolute GAG content (mg/mg) and water content of the probes. The calculated GAG concentrations were expressed as the relative weight per cartilage wet weight [% GAG/mg WWt]. To compare MRI data to biochemical analysis, cartilage areas were segmented in MR images and gagCEST values were averaged in regions corresponding to the division used for biochemical analysis. The correlation coefficient (r) for gagCEST and biochemical essay was determined using Pearson correlation analysis. To account for individual differences in cartilage water content, which can alter chemical exchange effects, measured gagCEST signals were scaled to the fictive case of 90 % water content in cartilage. Results: All examined ankles showed morphologically intact cartilage on PDw MR images. From the 42 available cartilage samples (7 patients x 2 cartilage surfaces x 3 cartilage segments 1⁄4 42), 4 samples from ankle # 6 were excluded from analysis due to extremely thin (<< 0.8 mm) cartilage in the medial and lateral segments. The remaining 38 data points showed a linear correlation between gagCEST signal intensities and GAG concentrations with r 1⁄4 0.797 if differences in water contents were neglected. If these differences were accounted for, a higher correlation coefficient of r 1⁄40.859 was obtained (Fig. 1b). The average measured gagCEST signal intensity (Fig. 1c) was 5.47 3.52 % (mean SD), and 8.11 5.32 % with normalized water content. The average GAG content as determined by biochemical analysis was 6.49 1.14 % GAG/mg WWt. The relative water content in cartilage had a mean of 68.85 4.54 %.