In-Vivo Measurement of Serine in Human Brain by Constant-TE PRESS Difference Editing at 7.0 Tesla

INTRODUCTION: Serine (Ser), a co-agonist at N-methyl-D-aspartate receptors, has three coupled spins resonating at 3.98, 3.94 and 3.83 ppm [1]. Because of its low concentration (~0.5 mM) and the overlap with primarily the creatine (Cr) 3.92 ppm singlet, detection of Ser in human brain by H-MRS remains as a challenge. A spectrally-selective refocusing approach, that is the first attempt to measure this experimentally-challenging metabolite, was reported a year ago [2], which employed a narrow-band 180° RF pulse to refocus the Ser 3.83-ppm resonance and to dephase the Cr signal and other neighboring resonances. The clinical utility of this method may not be very promising because the signal intensity and lineshape are both sensitive to the potential frequency drift, arising from subject motion during the long measurement time required to achieve an acceptable SNR in vivo. Here, we propose a constant-TE PRESS difference editing strategy for detection of Ser, which utilizes difference in spectral pattern of strongly-coupled spins at unequal subecho times (TE1 and TE2) with a single (total) TE. A preliminary in vivo result at 7T is presented. METHODS: Due to the coherence proliferation between coupled resonances and the imperfect excitation profile of volume selective RF pulses, the signal from strongly coupled spins is not fully described by the total echo time (TE) but also depend on the subecho times of a PRESS sequence. A difference in spectral patterns of Ser between unequal pairs of subecho times with an identical total TE can be utilized to reveal a Ser signal and cancel out the Cr 3.92-ppm singlet which depends on TE only. Subecho time dependence of the Ser multiplet at 7T was investigated for TE1 and TE2 between 20 – 200 ms with 1 ms increments, using density-matrix simulation incorporating the slice-selective shaped RF and gradient pulses. Spatial localization was obtained with an 8.8-ms 90° RF pulse (BW = 4.7 kHz) and two 11.9ms 180° RF pulses (BW = 1.4 kHz). In vivo validation of the simulation result was conducted at 7T (Philips Medical Systems, Cleveland, OH, USA), using a quadrature birdcage head RF coil with 16 reception channels for RF transmission and reception. A 40×40×40 mm voxel was positioned in the medial prefrontal cortex of a healthy volunteer (see Fig. 3).