Fluid filling of the digestive tract for improved proton resonance frequency shift-based MR thermometry in the pancreas

Purpose: To demonstrate that fluid filling of the digestive tract improves the performance of respiratory motion-compensated proton resonance frequency shift (PRFS)-based magnetic resonance (MR) thermometry in the pancreas. Materials and Methods: In seven volunteers (without heating), we evaluated PRFS thermometry in the pancreas with and without filling of the surrounding digestive tract. All data acquisition was performed at 1.5T, then all datasets were analyzed and compared with three different PRFS respiratory motion-compensated thermometry methods: gating, multibaseline, and referenceless. The temperature precision of the different methods was evaluated by assessing temperature standard deviation over time, while a simulation experiment was used to study the accuracy of the methods. Results: Without fluid intake, errors in temperature precision in the pancreas up to 10°C were observed for all evaluated methods. After liquid intake, temperature precision improved to median values between 1.8 and 2.9°C. The simulations showed that gating had the lowest accuracy, with errors up to 7°C. Multibaseline and referenceless thermometry performed better, with a median error in the pancreas between –3 and +3°C after fluid intake, for all volunteers. Conclusion: Preparation of the digestive tract near the pancreas by filling it with fluid improved MR thermometry precision and accuracy for all common respiratory motion-compensated methods evaluated. These improvements are attributed to reducing field inhomogeneity in the pancreas. Level of Evidence: 2. Technical Efficacy: Stage 1. J. Magn. Reson. Imaging 2018;47:692–701.