Aim of the study: Using phantom studies, a dedicated low-field MR system with 0.2 T and a whole-body MR-scanner with 1.0 T were compared.
Methods: A spin-echo sequence was performed on the 0.2-T MR unit using the knee coil and on the 1.0-T MR unit with the head coil. In a water-filled phantom, signal-to-noise ratios (SNR) were calculated and contrast measurements on gels with well-defined relaxation times were obtained and compared to nominal relaxation times. Measurements of T1 and T2 relaxation times on the low-field system were compared to the 1.0-T unit. As a parameter for geometrical image quality, magnetic field distortions were calculated. In theory, influence of field strength and and receiver bandwidth on the minimal echo time and on chemical shift artifacts were calculated.
Results: The SNR was 63.2 on the 0.2-T and 179.6 on the 1.0-T MR unit (difference factor 3, against a theoretical difference of 5-16). Relaxation times on the low-field system were significantly (around 40-50%) lower. Measurements of contrast were similar on both systems. On the low-field system geometrical distortions of several pixels were recorded. The minimal echo time on the low-field system was 21.3 ms with the low receiver bandwidth and 4.3 ms on the 1.0-T MR unit. The amount of chemical shift artefacts was the same on both systems.
Discussion: On low-field MR systems SNR is markedly improved by small read-out gradients and a low receiver bandwidth (factor 3 vs 5-16 in comparison with the 1.0-T MR unit). However, an optimal homogeneous magnetic field is required, since inhomogeneities may create severe geometrical distortion.