Background

Proton magnetic resonance spectroscopic imaging (MRSI) of the prostate has been used with only a combination of external surface coils. The quality of spectral fitting of the (choline + creatine)/citrate ([Cho + Cr]/Cit) ratio at different field strengths and different coils is important for quantitative/semi-quantitative diagnosis of prostate cancer.

Purpose

To evaluate the quality of spectral fitting of the (Cho + Cr)/Cit ratio of a prostate phantom using MRSI at different field strengths and various coils.

Material and Methods

Experiments were using 1.5-T and 3.0-T MR systems. Measurements were taken on a homemade prostate phantom with different coils: spinal array; abdominal array; and endorectal. The signal-to-noise ratio (SNR) of choline, creatine, and citrate peaks as well as the (Cho + Cr)/Cit ratio in each voxel were compared among groups using multi-way analysis of variance.

Results

Magnetic field strength, coils, and plane position had a significant effect on the SNR or (Cho + Cr)/Cit ratio, and there were interactions among groups (all P = 0.000). The 1.5-T (0.228 ± 0.044) exhibited a higher (Cho + Cr)/Cit ratio than the 3.0-T (0.125 ± 0.041) magnetic field strength (F = 3238, P = 0.000). The (Cho + Cr)/Cit ratio of both surface coils (0.183 ± 0.065) and all coils (0.181 ± 0.057) was significantly lower than that of the endorectal coil (0.195 ± 0.077) (both P < 0.05), but significant differences in the mean (Cho + Cr)/Cit ratio were not observed if surface coils and all coils were used (P > 0.05). No significant differences were found among the (Cho + Cr)/Cit ratios of all voxels in the middle planes by the post-hoc analyses (all P > 0.05).

Conclusion

Three-dimensional proton MRSI of prostate metabolites in a phantom using surface coils is feasible and reliable, but (Cho + Cr)/Cit ratios acquired at different magnetic fields and coils were different. This difference should be taken into account when calculating this ratio in a field strength-independent way.