mapping of the brain from water-unsuppressed ¹ H-MRSI and turbo spin-echo data

To obtain high-quality $$ maps of brain tissues from water-unsuppressed magnetic resonance spectroscopic imaging (MRSI) and turbo spin-echo (TSE) data.

$$ mapping can be achieved using $$ mapping from water-unsuppressed MRSI data and $$ mapping from TSE data. However, $$ mapping often suffers from signal dephasing and distortions caused by $$ field inhomogeneity; $$ measurements may be biased due to system imperfections, especially for $$ -weighted image with small number of TEs. In this work, we corrected the $$ field inhomogeneity effect on $$ mapping using a subspace model-based method, incorporating pre-learned spectral basis functions of the water signals. $$ estimation bias was corrected using a TE-adjustment method, which modeled the deviation between measured and reference $$ decays as TE shifts.

In vivo experiments were performed to evaluate the performance of the proposed method. High-quality $$ maps were obtained in the presence of large field inhomogeneity in the prefrontal cortex. Bias in $$ measurements obtained from TSE data was effectively reduced. Based on the $$ and $$ measurements produced by the proposed method, high-quality $$ maps were obtained, along with neurometabolite maps, from MRSI and TSE data that were acquired in about 9 min. The results obtained from acute stroke and glioma patients demonstrated the feasibility of the proposed method in the clinical setting.

High-quality $$ maps can be obtained from water-unsuppressed ¹ H-MRSI and TSE data using the proposed method. With further development, this method may lay a foundation for simultaneously imaging oxygenation and neurometabolic alterations of brain disorders.