Abstract
PURPOSE: To evaluate the feasibility of replacing clinically utilized 2D thermometry with 3D segmented EPI-based thermometry with equivalent accuracy, precision, and scan time. METHODS: A 3D segmented EPI (segEPI) trajectory was modified to allow readouts along both the forward and reverse direction for each phase encoding line, and the sequence was integrated into the existing clinical protocol. Focused ultrasound sonications were performed in a tissue-mimicking gelatin phantom with the clinical standard scan and the proposed scan across multiple repetitions. Accuracy, precision and extent of the measured heating was compared at varied levels of zero-filled interpolation. Non-heating in vivo brain scans were performed to evaluate phase drift and precision improvements from echo combination. RESULTS: A 3D segEPI thermometry was found to have high agreement with 2D clinical thermometry while providing greater knowledge of the extent of heating. Temperature limits of agreement remained below ±1°C for both the uncombined and combined-echo cases when zero-filled interpolation was performed to 0.5 mm in-plane voxel spacing or smaller. The full-width at half maximum of the heating distribution was found to be slightly higher (∼17%) for 3D segEPI, potentially due to uncorrected heating-induced chemical shift. Echo combination in segEPI is found to improve precision of temperature estimates. CONCLUSIONS: A 3D segEPI thermometry is shown to be a promising alternative to clinical 2D thermometry, with possibility of rapid integration into existing clinical workflows.