Qing Dai

Purpose: To develop a software-based active electromagnetic interference (EMI) suppression (AES) framework on whole-body multi-channel MRI systems for reliable MRI and MR thermometry during MRI-guided microwave ablation (MWA). Methods: In addition to primary imaging coils, an unloaded standard body array coil was positioned outside the imaging volume to serve as "auxiliary coils" for detecting and characterizing EMI. At each time frame, principal component analysis compressed the auxiliary coil data to extract the dominant EMI pattern. The EMI coupled into the primary coils was then modeled with channel-specific impulse response functions and subtracted from the primary coil data on a frame-by-frame basis. The framework was evaluated through MWA experiments in gel phantoms and in vivo pig models at 3T with parallel imaging acceleration. Proton resonance frequency shift (PRF) MR thermometry accuracy was assessed using fiber-optic probes. The PRF-derived thermal dose prediction was compared with post-ablation MRI and tissue gross pathology. Results: AES effectively suppressed EMI-induced artifacts and substantially improved image quality, yielding signal-to-noise-ratio enhancement of 40-fold (gel) and 13-fold (in vivo) and an EMI suppression rate >92%. Reliable PRF temperature mapping was enabled, with temporal mean-absolute-error of <1.4°C and <0.3°C in ablated and non-heating regions, respectively. Thermal dose estimation matched post-ablation MRI and gross pathology findings. Conclusion: The proposed AES framework effectively suppressed ablation-device-related EMI and enabled reliable MRI and MR thermometry during MWA. This approach can be integrated into existing reconstruction pipelines to facilitate the broader clinical adoption of MRI-guided MWA and other interventions affected by EMI.