A Simplified Procedure for Remote Tuning Local MRI Coils for Maximum Signal-to-Noise Ratio

INTRODUCTION: The purpose of this work is to develop a simple method to achieve the best signal-to-noise ratio (SNR) for a surface coil without the required use of a network analyzer and do so when the coil is under load inside the magnet. In our lab, before imaging, a local coil was connected to the S11 input of a network analyzer and tuned to the resonance at 50 ohm impedance. When remote voltage controlled adjustment was implemented it was found that we can obtain better SNR by a slightly readjusting the coil impedance from initial conditions. It was first found for a cryogenic coil that showed lower background noise on the image than the same coil before cooling. The results presented here were obtained for coil-preamplifier systems without a cryogenic cooling. METHODS: The study was performed on a GE Signa EXCITE 3T MR scanner. Acquisition was done off-line using a computer equipped with Mercury Computer Systems, Inc. (Chelmsford, MA) ECDR-GC316 cards running Linux OS (1). A tune/match circuit was built as part of a surface coil, 28 mm in diameter, using non-magnetic parts, including two GaAs hyperabrupt varactor diodes MGV125-25. The box with two dials and two digital voltmeters was built to control voltages supplied to varactors. Dials marked “match” and “tune” were used to adjust a coil impedance and resonance. Digital readouts were invaluable when a return to initial conditions was required. The Hewlett Packard RF network analyzer 8712ES was used to measure the coil impedance on a Smith diagram. Four different preamplifiers were used in the experiment. A GE 3T scanner internal amplifier was used as a reference. In addition, we tested preamplifiers from (A) USA Instruments P/N 1101457 with a gain of 27 dB, (B) Sirenza on evaluation board containing an SPF-5122Z transistor with a gain of 25 dB, and (C) WanTcom, model WMA3RA (Burnsville, MN) with a 30 dB gain. The input impedance of these amplifiers differed but had negligibly small reactance. Resistance was measured at 4 Ohm for A, 28 Ohm for B, and 0.8 Ohm for C units. Relative SNR measurements were done using the gradient recalled sequence: BW = 31.25 kHz; 256 × 256 resolution; FOV = 12 cm; slice = 1 mm; TR = 250 ms; TE = 7.2 ms, in the prescan mode as displayed in Fig. 1. Noise standard deviation was derived from I and Q data inside red-shaded areas by the formula