Assessment of bound pool fractions in the aging brain with stimulated echoes

Introduction Besides degeneration of axons, normal aging of human cerebral white matter (WM) is histologically associated with changes in myelin sheaths [1,2]. Exchange of magnetization between protons of tissue water and those bound to macromolecules is used in MRI to determine the bound pool fraction (BPF). BPF, the molar fraction of bound protons is associated with intact myelin and was previously suggested as a predictor of myelin content in postmortem white matter [3]. A recently presented single-shot EPI (sshEPI) imaging sequence, which is based on stimulated echo amplitude modulation (STEAM) allows for fast quantification of BPF [4]. The multislice method enables coverage of central brain areas in five minutes and is low in SAR, and therefore applicable at higher field strengths. Consequently, the STEAM approach seems promising for BPF mapping in clinical routine. To determine normative BPF ranges of the human brain, we conducted a pilot study in healthy volunteers of different age. BPF of five different WM brain structures and its difference were evaluated for two age groups of ten subjects each. Subjects and methods Study population consisted of one group of 10 young subjects aged between 31 and 44 years (5 male / 5 female, mean age 31y) and a second group of 10 subjects in the range of 47 to 77 years (4 male / 6 female, age: mean age 60y). A total of 23 subjects underwent MRI, of which 3 were excluded from further analysis due to motion artifacts. Measurements were performed on a 3T Tim Trio (Siemens Healthcare, Erlangen, Germany). As described in [4], the applied sequence consists of a tagging preparation followed by a multiple sshEPI scheme (Figure 1). Slice permutation enables imaging of multiple mixing times (TM) for a single tagging preparation. The basic experiment is repeated n times permuting the slice order such, that all TMs are acquired for each slice. Imaging parameters were: TM from 3.2ms to 603.2ms with 60ms TM-spacing, echo time of the EPI readout was 22ms, TR=2600ms, spatial resolution: 2.5x2.5x5 mm, slice gap: 20%, 11 slices, 11 experiments, NSA=10. Measurements TM2–TM10 were used for monoexponential fitting [4]. Signal acquired with minimal TM was assumed for net magnetization before any magnetization transfer occurred. Total acquisition time for covering a FOV of 25x25x6.6 cm was 4.8 minutes. T1 maps, which are obtained automatically as a by-product of the BPF estimation procedure, were used to manually outline regions of interest including splenium and genu of the corpus callosum, and corona radiata, frontal and occipital white matter were separately assessed for each hemisphere. To evaluate the significance of differences between the two age groups a sampled t-test, in which p<0.01 was considered to be statistically significant, was applied. Results Mean regional BPFs (n=10) for the two age groups (Table 1) were in good agreement with published data using other methods [5,6]. Standard deviations of both age groups were in the range of recently presented in-vivo BPF measurements at 3T in 5 subjects [5] but higher than reported for a balanced stead state free precession (bSSFP) approach at 1.5T in 12 subjects [6]. Except for frontal WM, all WM regions evaluated had significantly lower mean BPFs for the older age group (Table 1).