Motivation: MRI-guided percutaneous needle interventions offer high precision for targeted procedures, however, accurate needle localization remains a challenge due to limited real-time tracking capabilities. Goal(s): We developed a pysics-driven approach for needle localization in undersampled data. Approach: Needle artifacts were simulated for a golden-angle radial GRE sequence and used to develop a library-based localization method. Results: Accurate and robust localization was obtained using only 16 spokes, despite the appearance of strong streaking artifacts. Impact: Our approach shows potential to enable smoother guidance and increased accuracy during percutaneous interventions.
Schwieters et al. (Tue,) studied this question.