Highly Reproducible in vivo T 1 Maps in brain at 3 T

Purpose: Many factors affect the accuracy of quantitative MRI methods. For T1 mapping with a spoiled-gradient echo sequence, accuracy has been shown to be affected by B1 inhomogeneity, slice profile effects, the effectiveness of spoiling and noise. Some of these factors may vary spatially across the brain, so it is important to assess the effects of these factors in anatomically defined regions-ofinterest (ROIs). Furthermore, many of these factors may vary from scan to scan due to subject placement, field inhomogeneity, scanner instability, physiological noise and/or true physiological changes. It is essential to assess the reproducibility of a method when attempting to use the resulting metric to detect differences across subjects (eg. healthy vs diseased) or within subjects (eg. drug induced effects, aging). An efficient method of 3D B1-corrected T1 mapping has been proposed, the Method of Slopes (MoS), and recent work demonstrated the accuracy of the T1 maps in vivo. However, the reproducibility of the resulting T1 maps has not yet been determined. In this work, we assess the regionally-dependent reproducibility of T1 mapping with the MoS. Our goal is to quantify the intra-subject variability within a day (morning and afternoon) and across two days (subsequent mornings). Methods: Five healthy controls were recruited (ages: 35.6±9 years, 2 male) and consent was obtained according to the REB of the Institution. Subjects were scanned with a 3T scanner (MR750, GE Healthcare) using the fastest protocol described in ref.5, yielding B1-corrected T1 maps with an isotropic resolution of 1mm in less than 10 minutes. Subjects were scanned at the three time points: the morning and afternoon of the same day and the morning of the following day. Resulting T1 maps were co-registered within subject, using a linear registration algorithm in FSL (FMRIB Analysis Group, Oxford University, UK). 118 ROIs were identified using the AAL template. T1 values were extracted per ROI. Histograms were plotted to show the distribution of T1values per ROI and an average value was calculated at each time point i=1,2,3: 〈T1〉i (for each subject and ROI). Results: Fig.1 illustrates the mean ± std of 〈T1〉i for i=1,2,3 in some representative ROIs. The intra-subject variability is indicated by the size of the error bars. The inter-subject variability is reflected in the variation between the height of the different-colored bars (per ROI). The reproducibility was quantified for all ROIs as a coefficient of variation (CV) in time for intra-subject variability and across subjects for inter-subject variability. Results, in all ROIs, gave CV<8% and CV<12% for intra-subject and inter-subject variability, respectively.