Dedicated two-channel phased array receiver coils for HR-MRI of the rat knee cartilages at 7T

2 internal and 18 x 20 mm 2 external dimensions). One conductor of this loop is common for the two elements. The S-parameters and the quality factor were measured with an ENA300 network analyzer (Agilent Technologies Inc., Santa Clara, CA, USA). The decoupling between the two channels was achieved using a fixed capacitor inserted in the common conductor. The value determined by simulation was experimentally adjusted to minimize the S21 transfer parameter between the two channels. Both channels were tuned at 300.3 MHz corresponding to the proton's resonance frequency at 7 T and matched to 50 impedance line using non-magnetic case A series 100 and 710 ATC capacitors (American Technical Ceramics, New York, USA). The tuning/matching and active decoupling circuit was designed to be interfaced with the system decoupling box. The MRI experiments were performed on a 7T Biospec (Bruker, Ettlingen, Germany) equipped with 4-proton receiver channels. The designed phased array coil was compared to a Bruker 15 mm diameter surface coil. Experimental characterization (signal uniformity and SNR) was performed on cylindrical phantoms filled with salty water (NaCl 0.45%) mimicking load conditions. For in vivo experiments, the ethical guidelines for experimental investigations with animals were followed, and the experimental protocol was approved by the Animal Ethics Committee of our institution. Gaseous anesthesia was performed on adult rats placed in supine position. The 15 mm diameter surface coil was placed in contact with the medial side of knee joint and the designed array coil was placed on top of patella to encompass the whole knee joint. The HR-MRI of the rat knee joint was performed using a 3D Gradient-Echo Fast Imaging (GEFI) sequence with the following parameters: 25° flip angle, 50 ms TR, 3.4 ms TE, 42 kHz rbw. A total of 64 partitions (312 µ m thick) were acquired with a FOV of 30 x 30 mm 2 and an acquisition matrix size of 512 x 384. Acquisition volume was reconstructed to a 512 x 512 x 128 matrix leading to a 156 µ m partition thickness and an in-plane pixel of 59 x 59 µ m 2 . The scan time for the GEFI sequence was 45 min. The femoral and tibial plateaus (medial and lateral) articular cartilage volumes were extracted using an interactive touch-sensitive screen with a 1280 × 1024 pixel matrix. The user segmented the knee cartilage compartments directly on this screen using the supplied pen. The articular cartilage was segmented on each MRI slice. Each segmented area was then assigned to its corresponding cartilage compartment using gray-scale code labels leading to the three articular cartilage volumes. Additionally, a proof of concept was performed using two independent phased array coils (one for each knee joint). The phased array coils were decoupled using a 20 x 25 mm 2 copper sheet, placed at equal distance between the two array coils. Additional supply voltage sources were used for tuning and matching of the second array coil. Both legs were acquired within the same scan using a similar acquisition as for a single knee joint with the same pixel size but with a larger FOV in a coronal plane. Results The measured quality factor of the unloaded coil was about 130 for every single channel. The quality factor of the loaded coil decreased to 110. The decoupling capacitor value mounted on the circuit was about 56 pF corresponding to a simulated mutual inductance of 9.4 nH. The isolation between the two channels was 27 dB. The decoupling between the two array coils separated by a thin copper layer for multiple knee imaging was 28 dB. The SNR gain in the ROI for the two-channel array coil was up to 2.2 compared to the SNR obtained with the 15 mm diameter surface coil. The signal intensity was more uniform with a SNR standard deviation of 5 compared to 27 measured on images acquired with the surface coil. The 2.2 gain in SNR was used in vivo to decrease the voxel size from 59 x 59 x 156 µ m 3