Sheng Chen, Steven P Allen, John P Mugler, G Wilson Miller, Craig H Meyer
The proposed method enabled inter-sonication monitoring of skull temperature changes and brain thermometry during transcranial MRgFUS. While accurate ex vivo, improvements to in vivo T1 precision are needed for clinical temperature quantification.
PURPOSE: To develop an MRI approach for simultaneous skull and brain temperature monitoring between sonications in transcranial MR-guided focused ultrasound (MRgFUS) surgeries.
METHODS: A 3D stack-of-spirals dual-echo ultra-short TE (UTE) sequence incorporating fat suppression, spiral deblurring, trajectory measurement, and B1 correction was implemented. A variable TE strategy enabled UTEs as short as 0.05 ms for skull T1 mapping, while a fixed late-TE (LTE) was used for brain temperature measurement. The approach was evaluated in laboratory and clinical settings through phantom and human studies.
RESULTS: During laboratory phantom cooling, MR-estimated temperatures closely matched optical fiber measurements, with standard deviations of 0.60°C and 0.57°C across two gel ROIs. Cortical bone T1 showed a strong correlation with temperature (Pearson r = 0.963, p < 0.05; slope = 2.51 ms/°C). Similar results were obtained during laboratory phantom heating (r = 0.800, p < 0.05; slope = 3.86 ms/°C). In clinical phantom heating, gel temperature and bone T1 changed concurrently, with peak pixel-wise temperature increases of ∼12°C and peak ROI-averaged T1 increases of ∼14 ms. In patient scans, B1 correction reduced skull T1 spatial standard deviation from 98.1 to 59.9 ms. Volunteer uncertainty maps showed mean uncertainties of 46.8 ms for skull T1, enabling detection of ∼30 ms changes with 80% power, and 0.57°C for brain temperature.
CONCLUSIONS: The proposed method enabled inter-sonication monitoring of skull temperature changes and brain thermometry during transcranial MRgFUS. While accurate ex vivo, improvements to in vivo T1 precision are needed for clinical temperature quantification.