Fast quantitative T2 mapping at ultra-high field using sparse sampling and dictionary-based reconstruction

Quantitative T2 mapping at ultra high-field via multi-echo spin-echo acquisitions suffers from bias introduced by pronounced transmit field inhomogeneities. Hence, accurate signal modeling is vital. Futhermore, high-resolution imaging results in longer scan times, which need to be compensated by k-space undersampling. Conventional undersampling schemes for multi-echo approaches skip identical phase encoding areas for all echo readouts. Image reconstruction quality might be improved by varying the sampling across echoes and using a low rank reconstruction approach. A multi-echo acquisition combining sparse undersampling with a dictionary modelling method was successfully used to acquire accurate T2 at 7T with a high acceleration.