Purpose

MR-Thermometry (MRTh) is key for guidance and supervision of thermal therapies^1-3 and beneficial for monitoring and improving MR safety^4-5. Proton resonance frequency (PRF) shift is commonly used for MRTh of aqueous tissue. For MRTh of adipose tissue temperature dependence of T₁ and T₂^6-7 relaxation times is exploited. Hybrid methods that allow simultaneous acquisition of water and fat based thermometry are of great clinical need⁸. Recognizing this need this work proposes a hybrid temperature mapping technique, designated as 2in1-Thermometry, which combines simultaneous PRF and T₂-based temperature mapping using simultaneous dual contrast weighting with double echo RARE imaging. 2in1-Thermometry is carefully validated in phantom heating experiments and benchmarked against conventional gradient echo (GRE) based PRF thermometry. Preliminary results of RARE based ΔT₂ thermometry in bovine suet are validated against fiber optic temperature sensor readings.

Methods

2in1-RARE imaging supports two contrasts in one acquisition by independent sensitization of stimulated echoes (STE) and spin echoes (SE)⁹. For PRF-MRTh an evolution time is incorporated between the initial excitation and the first refocusing pulse to generate T₂*-weighting for the SE magnetization¹⁰. The STE-image is not affected by this evolution time and is acquired for two TEs permitting simultaneous T₂-based MRTh assuming a mono-exponential T₂-decay for fatty tissue11 (Fig.1c). Radiofrequency (RF) heating was performed at f=300MHz using a bow tie dipole-antenna fed with an external RF-power amplifier¹² (Fig.1b). MR images were acquired at 3.0T¹³. To examine the fidelity of PRF-MRTh of 2in1-RARE, heating experiments were performed in phantoms and benchmarked against GRE based PRF. For this purpose a uniform agarose phantom was heated four times for three minutes (P_avg=70W at the antenna). B₀-drift correction was performed using oil samples placed around the phantom. To study the feasibility of T₂ based MRTh, heating experiments of fatty tissue were performed using conventional RARE and split-echo RARE^11,14 as a precursor of 2in1-RARE. For this purpose an agarose phantom including an insert containing bovine suet was heated for 6x6min (P_avg=70W at the antenna). Before heating and directly after each heating period RARE and split-echo RARE were performed to calculate a ΔT₂-map based on a mono-exponential decay (TE₁=17ms, TE₂=52ms). Temperature maps were calculated based on T₂ changes versus the reference image. MR-temperature maps were compared to measurements of fiber optic sensors¹⁵ placed inside the phantom.

Results

PRF temperature maps calculated from 2in1-RARE and GRE are shown in Fig.2. The match between the maps demonstrates the feasibility of PRF based thermometry of 2in1-RARE using a slightly longer acquisition time to ensure a similar SNR. The temperature rise over time is displayed in Fig.3 for one fiber optic sensor position. The difference between the fiber optics data and the MR based temperature averaged in a region of interest (ROI) around the sensor position was defined as the accuracy (dT) of the temperature mapping technique. These accuracies were averaged for all four sensor positions and all MR measurements for a final accuracy of dT=(0.5±0.2)°C for GRE based PRF-MRTh and dT=(0.6±0.2)°C for 2in1-RARE based PRF-MRTh. The temperature dependent ΔT₂ measurements of split-echo RARE in bovine suet tissue were comparable to ΔT2 values acquired using RARE (Fig.4). In comparison to fiber optic sensor readings in the bovine suet sample, RARE based temperature dependent ΔT₂ changes showed a linear dependency with ΔT changes¹¹ (Fig.4d).

Discussion

The presented data demonstrate that 2in1-RARE based PRF thermometry is feasible and supports a temperature mapping fidelity competitive with GRE based PRF-MRTh. To take our developments to the next level the presented preliminary T­₂ based thermometry for RARE and spilt-echo RARE will be incorporated into 2in1-RARE. However the echo separation in 2in1-RARE leads to signal dephasing affected not only by T₂ dephasing and consequently smaller T₂ changes compared to RARE are presumed. A T₂ preparation module¹⁶ will be added to the 2in1-RARE sequence and is expected to produce similar results to RARE and split-echo RARE, which showed good linearity (ΔT₂/ΔT) as compared to fiber optic readings. The proposed 2in1-RARE-MRTh is especially beneficial for temperature monitoring applications where the traditional gradient echo and echo planar imaging based approaches are compromised bysusceptibility artifacts and image distortion. For the proposed 2in1-RARE-MRTh approach temperature sensitization can be varied from zero, which is not feasible with conventional gradient echo based PRF-MRTh.

Conclusion

A spin echo based pulse sequence - 2in1-RARE thermometry – was presented that offers the potential to simultaneously acquire two images with different contrasts (T₂* and T₂). This enables the simultaneous calculation of temperature maps in adipose and water-based tissue which is a remaining challenge for GRE based thermometry.

Acknowledgements

This work was supported in part (L.W., T.N.) by the German Federal Ministry of Education and Research, “KMU-innovativ”: Medizintechnik 13GW0102.