METHOD for CENTER FREQUENCY CALIBRATION USING STIR ON/OFF SPECTRA

Motivation: To propose a robust method for center frequency (F0) calibration. Goal(s): To improve image quality across a wide range of clinical conditions by improving the accuracy and robustness of center frequency calibration. Approach: A model-based peak fitting approach is applied to spectra acquired with STIR fat suppression on and off. The model kernels are tuned individually for STIR on/off data. By jointly solving F0 in both STIR on/off data the F0 calibration is robust even in the presence of strong fat signal (e.g. breast). Results: The proposed method was successful in 100% of test cases (51/51) across a range of human anatomies. Impact: Accurate and consistent center frequency calibration is critical for reliable image quality. The proposed method is robust to a wide range of clinical conditions and anatomies, including tissues with high off-resonance (e.g. cervical spine) or high fat content (e.g. breast).