Automatic Varactor Tuning of Interventional RF Receiver Coils R. Venook 1 , G. Gold 1 , B. Hu 2 , G. Scott 1 1 Stanford University, Stanford, CA, United States, 2 Palo Alto Medical Foundation, Palo Alto, CA, United States Synopsis: Small, flexible RF coils are well suited for interventional applications because of their locally-high SNR and deployability. These characteristics also make such coils easily detuned/mismatched, which can hamper imaging performance. This work demonstrates an autotuning RF receive system that functions in a 1.5T GE scanner, and successfully improves SNR by ~70% with a tuning time under 1 second. Introduction and Background: Many interventional MRI applications are enabled by the high local SNR of small receive coils [1]. However, greater coupling to the tissue of interest also creates heightened sensitivity of these coils to local environment. Moreover, while flexible coils are most compliant to the constraints of in vivo use, changes in coil shape can ruin their tuning as well. Mis- tuning of a coil degrades its match to the preamplifier, which lowers system SNR by raising the noise figure of the receive chain [2, 3]. In lower-field interventional systems, the sensitivity will be even greater due to higher loaded coil Q’s. Hence, tuning and matching are required to maximize the benefit of interventional imaging. Unfortunately, manual tuning and matching is time-consuming and difficult. For imaging under dynamic conditions, it is completely untenable. Prior work in automatic tuning and matching requires either difficult coil topologies, or mechanical tuning devices, or both [4]. We have previously proposed a system that uses a microcontroller, phase detector and a varactor-diode (as a voltage-controlled tuning capacitor) to automatically tune the RF receive coil via DC bias at the press of a button (figure 1b) [5]. Here we present the results of phantom imaging experiments using an improved, MR-compatible version of that system. Methods: Electronics—Much of the autotuning electronics in this work are improved variations on our previous scheme. Modifications were made to improve system robustness and allow operation within a clinical scanner. The phase detector was rebuilt with surface-mount multiplier and comparators to reduce power drain and improve tuning stability; and, a simple, linear power supply was built to allow operation with only two 9-volt batteries. Testing—A cylindrical phantom was imaged with our flexible, 2cm diameter, double-loop, varactor-tuned, receive-only RF coil [6] in different coil configurations, and with different tuning mechanisms (automatic vs. passive). A GE Signa 1.5T scanner was used, with a stock SPGR sequence (TE=7.2ms, TR=34ms, Res.=0.47mm 2 , 256x256, 4mm slice, 30º tip). First, the coil was auto-tuned using our circuitry and imaged. Then, the coil’s shape was made narrower (by 50% in the into-the-plane direction) and an image was taken before retuning. Then, after re-autotuning, a third image was acquired. This process was repeated for a 50% coil width expansion. Between all stages an MR-compatible DIP meter was used to verify tuning condition. SNR profiles were calculated to quantitatively compare imaging performance under different coil shape and tuning conditions. Results and Discussion: The result of our electronic modifications is MR-compatible autotuning circuitry that repeatably tunes a coil to the Larmor frequency in a 1.5T scanner (figure 1a). Changing the coil width by 50% (narrower and wider, respectively) detuned the coil by –4 and 2.5 MHz. Autotuning after the shape changes created ~70% improvement in SNR (figure 2, bottom). This improvement is seen in the images of figure 2 (a-c). Also, SNR comparison of these autotuned images with images obtained using a passively tuned version of our receiver yielded no significant difference, meaning that the autotuning circuitry does not cause SNR degradation. Finally, note that the peak SNR in the initial and final images are expected to be different due to differences in the tuned impedance of the coil with different loop shapes. This mis-match is responsible for ~12% tuned SNR variation for the setup here, but for higher-Q receivers matching correction may be necessary. Interventional MRI applications have become increasingly popular, and as interventionalists strive for better SNR they will need ways to automatically tune and match their receivers. This approach will allow convenient and dynamic tuning of deployable catheter coils for real SNR advantage. References [1] Gold, G. et al, Proc. ISMRM 2001; 1: 84. [2] Hoult, D. et al, JMR 1976; 24: 71-85. [3] Fish, P. Elec. Noise and Low Noise Design: Ch. 5. McGraw Hill. [4] Hwang, F. et al, MRM 1998; 39(2): 214-22. [5] Venook, R. et al, Proc. ISMRM 2002: 893. [6] Scott, G. et al, Proc. ISMRM 2001; 1: 20. Figure 1: (a) Picture of the autotuning circuit, RF coil, and phantom; operation in the magnet bore. (b) Block diagram of autotuning circuitry (b) (a) (a) (b) (c) S N R Image Pixel Number Figure 2: Phantom images with (left to right) autotuned skinny coil, widened coil not re-tuned, and re-autotuned widened coil. (a-c) are windowed with equal noise power to show signal loss when not autotuned (b). The bottom plots show pixel-wise SNR for a central, horizontal line through each image. Tuned De-tuned Re-tuned 686 Proc. Intl. Soc. Mag. Reson. Med. 11 (2003)