T-1 FAST ACQUISITION RELAXATION MAPPING (T-1-FARM) - AN OPTIMIZED RECONSTRUCTION

Maps of spin lattice relaxation time (T-1) can be reconstructed direct ly from magnetic resonance imaging (MRI) k-space data measured with ve ry short data acquisition times, e,g,, a data set for a 128 x 128 T-1 map can be acquired in less than 3 s using gradients with 10-T/m/s sle w rate, In principle, this approach could be extended to quantitate ot her MRI parameters but current use is limited by the lack of precise, accurate and fast reconstruction. Using theoretical calculations, comp uter simulations, and experiments we have optimized a parametric recon struction method using a Leverberg-Marquardt (L-M) algorithm and compa red it to the quasi-Newton method originally used. We have found signi ficant improvement using the L-M method provided T-1 is solved for dir ectly without linearization, Reconstruction time was reduced by a fact or of 60. Computer simulations show that the method has acceptable acc uracy even in signals with 5% noise. Optimization included the investi gation of the signal-to-noise (S/N) of each k-space data point and its impact on relative error of the reconstruction. This result indicates that rectangular k-space data could be collected for further reductio n of data acquisition times. Determination of T-1 maps by direct param etric reconstruction of L-space data appears feasible and may stimulat e further application of quantitative MRI.