Fat-referenced MR thermometry in the breast and prostate using IDEAL

Purpose: To demonstrate a three-echo fat-referenced MR thermometry technique that estimates and corrects for time-varying phase disturbances in heterogeneous tissues. Materials and Methods: Fat protons do not exhibit a temperature-dependent frequency shift. Fat-referenced thermometry methods exploit this insensitivity and use the signal from fat to measure and correct for magnetic field disturbances. In this study, we present a fat-referenced method that uses interpolation of the fat signal to correct for phase disturbances in fat free regions. Phantom and ex vivo tissue cool-down experiments were performed to evaluate the accuracy of this method in the absence of motion. Non-heated in vivo imaging of the breast and prostate was performed to demonstrate measurement robustness in the presence of systemic and motion-induced field disturbances. Measurement accuracy of the method was compared to conventional proton resonance frequency shift MR thermometry. Results: In the ex vivo porcine tissue experiment, maximum measurement error of the fat-referenced method was reduced 42% from 3.3 to 1.9°C when compared to conventional MR thermometry. In the breasts, measurement errors were reduced by up to 70% from 6.4 to 1.9°C. Conclusion: Ex vivo and in vivo results show that the proposed method reduces measurement errors in the heterogeneous tissue experiments when compared to conventional MR thermometry. J. Magn. Reson. Imaging 2012;36:722-732.