Comparing 129Xe gas MRI across major MRI vendors and sites

Multi-site studies and clinical hyperpolarized 129Xe are imminent, yet image quality is challenging to quantify across MR platforms since hardware, gradient-corrections, and data-filtering are proprietary. We developed a straightforward protocol for scanning a commercial, thermally-polarized 129Xe phantom, and implemented it across three major MRI vendors, with near-identical coils. Images were acquired at 5 sites (Cincinnati, Iowa, Wisconsin, Duke, and Virginia). SNR was 18.4±1.6, with low variation in SNR between sites (13%), and vendors (10%). This confirms the ability to quantitatively compare signal and noise across sites and vendors. It is also demonstrated, here, that quality control can be implemented simply in multi-site studies.