Osteoarthritis (OA) is a degenerative disease of the whole joint that is characterized by early changes in articular cartilage. Mapping of changes in cartilage parametric values on MRI is helpful in early OA detection and monitoring of disease progression [1]. The T2 relaxation time and apparent diffusion coefficient (ADC) in cartilage is correlated with collagen matrix organization and hydration [2,3]. Quantitative Double-Echo Steady-State (DESS) MRI has been shown to be an efficient way to acquire 3D cartilage data including thickness, T2 relaxation times, and ADC [4-6]. Development of the DESS technique for 7T has the potential to improve quantitative accuracy and speed of the method. The current study compares T2 relaxation time and ADC values obtained from a set of healthy volunteers at 3T and 7T.

Acquisition

Healthy volunteers with no history of knee pain, injury, or surgery were imaged at both 3T and 7T (GE Healthcare, Milwaukee, WI). Ten subjects (7F, age 25-56) were imaged at 3T and six subjects (4F, age 25-56) at 7T. All scans were done under IRB approval. At 3T, imaging was done with a 16-channel flex coil (GE Healthcare, Milwaukee, WI) while at 7T it was performed with a 28-channel coil (QED, Mayfield Village, OH). Two DESS sequences were acquired at each field strength with TR/TE of 32/12 ms, 3 mm slice thickness, 256x256 matrix, 16 cm field-of-view, and imaging time of 5 min. A custom spectral-spatial pulse was used at each field strength for water excitation. Spoiler areas were 156.60 mT/m*ms and 15.66 mT/m*ms, giving high and low diffusion weighting, respectively.

Parametric Mapping

T2 and ADC values were estimated by combining the four DESS echoes into two signal ratios, which have good sensitivity to T2 and ADC but are mostly insensitive to T1 [7]. The ratios were compared to a dictionary of ratios for a reasonable range of T2 and ADC, computed using numerical extended phase graph modeling [8]. The parameters of the dictionary entry that best matched the measurements then gave an estimate at each pixel. B1 errors were accounted for by using B1 maps, acquired using the Bloch-Siegert method at 3T and the Double-Angle method at 7T, to assign a flip angle measurement to each pixel.

Data Analysis

Identical regions of interest were placed in the articular cartilage at similar locations for all images by an experience radiologist. We measured the T2 relaxation time and ADC in the central medial formal condyle, central medial tibia plateau, central lateral femoral condyle, and central lateral tibia plateau cartilage in each volunteer at each field strength. Mean relaxation times and ADC values for each field strength were calculated. The coefficient of variation (CV, standard deviation of the value divided by the mean) was calculated for each location for T2 and ADC at each field strength.

The T2 relaxation time showed a slight decrease at 7T compared with 3T, which has been observed previously [9]. The mean overall ADC at each cartilage location were similar across the two field strengths (Table 1). The CV for the T2 relaxation was slightly lower at 7T compared with 3T. The CV of the ADC measurement at each location was better at 7T compared with 3T, likely due to noise affecting the fit of this parameter (Table 2). At 3T, the T2 maps show the expected normal variations across the cartilage but the ADC maps show areas of poor fit (Figure 1). At 7T, the T2 maps and ADC are more uniform. The patella and trochlea cartilage showed areas of residual B0 inhomogeneity at 7T (Figure 2). The DESS method provides a convenient 3D approach to estimation of cartilage T2 and ADC at both 3T and 7T. DESS uses lower radiofrequency power at 7T than other methods. The measurement of ADC is improved at 7T because SNR is better, and fitting this parameter is particularly sensitive to noise. DESS at 7T shows more sensitivity to residual B0 inhomogeneity near the patella. Overall, implementation of DESS at 7T has the potential to improve the speed and accuracy of quantitative cartilage mapping.