Respiratory navigator-guided multi-slice free-breathing cardiac T1 mapping on a magnetic resonance-guided linear accelerator

BACKGROUND AND PURPOSE: Image-guided cardiac radioablation on a magnetic resonance-guided linear accelerator (MR-linac) is emerging as a non-invasive treatment alternative for patients with cardiac arrhythmia. Precise target identification is required for such treatments. However, owing to concerns with the use of gadolinium-based contrast agents during treatment with high-energy radiation, non-contrast alternatives must be considered. Native T1 mapping is a promising technique to delineate myocardial scar which can serve as a surrogate for the treatment target. Further, the likely presence of an implantable cardioverter defibrillator (ICD) in arrhythmia patients necessitates approaches that are robust to metal-related artefacts. MATERIALS AND METHODS: We implemented an electrocardiogram (ECG)-triggered free-breathing cardiac T1 mapping approach on an MR-linac, making use of a respiratory navigator to account for respiratory motion. The technique was validated in a motion phantom and tested in healthy volunteers. We also compared the use of different readout schemes to evaluate performance in the presence of an ICD. RESULTS: The free-breathing cardiac T1 mapping approach agreed within 5% compared with ground truth T1 in a motion phantom. In healthy volunteers, an average difference in T1 of -3.5% was seen between the free-breathing and breath-hold approaches, but T1 quantification was impacted by data discarded by the respiratory navigator. Compared to balanced SSFP, the spoiled gradient echo readout was much less susceptible to artefacts caused by an ICD, but the lower signal adversely affected T1 quantification. CONCLUSIONS: Free-breathing cardiac T1 mapping is feasible on an MR-linac. Further optimisation is required to reduce scan times and improve accuracy.