Evaluation of the keyhole technique applied to the proton resonance frequency method for magnetic resonance temperature imaging.

PURPOSE

To evaluate the temporal and spatial resolution of magnetic resonance (MR) temperature imaging when using the proton resonance frequency (PRF) method combined with the keyhole technique.

MATERIALS AND METHODS

Tissue-mimicking phantom and swine muscle tissue were microwave-heated by a coaxial slot antenna. For the sake of MR hardware safety, MR images were sequentially acquired after heating the subjects using a spoiled gradient (SPGR) pulse sequence. Reference raw (k-space) data were collected before heating the subjects. Keyhole temperature images were reconstructed from full k-space data synthesized by combining the peripheral phase-encoding part of the reference raw data and the center phase-encoding keyhole part of the time sequential raw data. Each keyhole image was analyzed with thermal error, and the signal-to-noise ratio (SNR) was compared with the self-reference (nonkeyhole) images according to the number of keyhole phase-encoding (keyhole-data size) portions.

RESULTS

In applied keyhole temperature images, smaller keyhole-data sizes led to more temperature error increases, but the SNR did not decreased comparably. Additionally, keyhole images with a keyhole-data size of <16 had significantly different temperatures compared with fully phase-encoded self-reference images (P < 0.05).

CONCLUSION

The keyhole technique combined with the PRF method improves temporal resolution and SNR in the measurement of the temperature in the deeper parts of body in real time.