Sparse regression-based delineation of air-motion artifacts for real-time correction of PRFS thermometry

Proton resonance frequency shift-based MR thermometry is widely used to non-invasively monitor thermal therapies in vivo. However, further clinical integration in deep hyperthermia is hampered by intestinal air-motion induced susceptibility artifacts. We developed a sparse regression approach to delineate susceptibility artifact sources. The resulting mask is then used to correct the artifact using existing methods from quantative susceptibility mapping. We verified our approach by a heated phantom experiment equipped with a moveable air volume and temperature probes. Here, we found a reduction in the mean absolute error from 1.6 degrees Celsius to 0.4 degrees Celsius, near the air-motion artifact.